WO2020108088A1 - 确定风力发电机组的塔架净空的方法和装置 - Google Patents

确定风力发电机组的塔架净空的方法和装置 Download PDF

Info

Publication number
WO2020108088A1
WO2020108088A1 PCT/CN2019/109391 CN2019109391W WO2020108088A1 WO 2020108088 A1 WO2020108088 A1 WO 2020108088A1 CN 2019109391 W CN2019109391 W CN 2019109391W WO 2020108088 A1 WO2020108088 A1 WO 2020108088A1
Authority
WO
WIPO (PCT)
Prior art keywords
tower
tip
blade
image
edge
Prior art date
Application number
PCT/CN2019/109391
Other languages
English (en)
French (fr)
Inventor
王百方
杨博宇
程庆阳
Original Assignee
北京金风科创风电设备有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京金风科创风电设备有限公司 filed Critical 北京金风科创风电设备有限公司
Priority to EP19890997.0A priority Critical patent/EP3744974A4/en
Priority to US16/976,403 priority patent/US11421659B2/en
Priority to AU2019390462A priority patent/AU2019390462B2/en
Publication of WO2020108088A1 publication Critical patent/WO2020108088A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F03—MACHINES OR ENGINES FOR LIQUIDS; WIND, SPRING, OR WEIGHT MOTORS; PRODUCING MECHANICAL POWER OR A REACTIVE PROPULSIVE THRUST, NOT OTHERWISE PROVIDED FOR
    • F03D—WIND MOTORS
    • F03D17/00—Monitoring or testing of wind motors, e.g. diagnostics
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2260/00—Function
    • F05B2260/80—Diagnostics
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/30—Control parameters, e.g. input parameters
    • F05B2270/33—Proximity of blade to tower
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05—INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05B—INDEXING SCHEME RELATING TO WIND, SPRING, WEIGHT, INERTIA OR LIKE MOTORS, TO MACHINES OR ENGINES FOR LIQUIDS COVERED BY SUBCLASSES F03B, F03D AND F03G
    • F05B2270/00—Control
    • F05B2270/80—Devices generating input signals, e.g. transducers, sensors, cameras or strain gauges
    • F05B2270/804—Optical devices
    • F05B2270/8041—Cameras
    • 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
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/72—Wind turbines with rotation axis in wind direction

Definitions

  • the present disclosure generally relates to the field of wind power technology, and more particularly, to a method and apparatus for determining the tower headroom of a wind turbine.
  • the clearance of the tower of the wind turbine refers to the distance from the tip of the blade to the surface of the tower during the rotation of the impeller.
  • the blade For wind turbines, if the blade sweeping occurs, the blade needs to be replaced, and the cost of a single blade is higher, which will increase the maintenance cost. At the same time, the unit needs to be shut down during the blade replacement, and the power generation will also be lost during the unit shutdown. Therefore, once the blade sweep tower occurs, it will bring greater economic losses to the wind farm.
  • the tower headroom of the wind turbine cannot be measured by a measuring tool, which results in the fact that the tower headroom of the wind turbine cannot be obtained in real time.
  • Exemplary embodiments of the present disclosure provide a method and apparatus for determining the tower headroom of a wind turbine to solve the technical problem that the tower headroom of a wind turbine cannot be measured in the prior art.
  • a method for determining the tower headroom of a wind turbine comprising: acquiring an image of the wind turbine during operation, the image including the tips of the blades of the wind turbine and the tower Barrel; determine the position of the tip of the blade of the wind turbine from the acquired image; identify the edge of the tower from the acquired image; calculate the tip of the blade to the determined position of the tip of the blade and the identified edge of the tower The distance of the edge of the tower to obtain the clearance of the tower.
  • an apparatus for determining the tower headroom of a wind turbine includes: an image acquisition module that acquires an image of the wind turbine during operation, the image including the wind turbine The tip of the blade and the tower; the tip detection module to determine the position of the tip of the blade of the wind turbine from the acquired image; the tower edge recognition module to identify the edge of the tower from the acquired image; the tower clearance determination module Based on the determined position of the tip of the blade and the identified edge of the tower, the distance from the tip of the blade to the edge of the tower is calculated to obtain the clearance of the tower.
  • a tower headroom monitoring system includes: an image capturer for capturing images of blades of a wind turbine during operation; and a processor configured to: Acquiring images including the tips of the blades of the wind turbine and the tower from the captured images; determining the positions of the tips of the blades of the wind turbine from the acquired images; identifying the edges of the tower from the acquired images; Based on the determined position of the tip of the blade and the identified edge of the tower, the distance from the tip of the blade to the edge of the tower is calculated to obtain the tower clearance.
  • a computer-readable storage medium storing a computer program, which when executed by a processor implements the above method for determining the tower headroom of a wind turbine.
  • the tower headroom of the wind power generator set can be determined in real time, so as to effectively avoid the occurrence of blade sweeping.
  • FIG. 1 shows a block diagram of a tower clearance monitoring system according to an exemplary embodiment of the present disclosure
  • FIG. 2 shows a flowchart of a method for determining a tower headroom of a wind turbine according to an exemplary embodiment of the present disclosure
  • FIG. 3 shows a schematic diagram of the installation position of the image capturer according to the first exemplary embodiment of the present disclosure
  • FIG. 4 shows a schematic diagram of an image captured by the image capturer according to the first exemplary embodiment of the present disclosure
  • FIG. 5 shows a schematic diagram of the installation position of the image capturer according to the second exemplary embodiment of the present disclosure
  • FIG. 6 shows a schematic diagram of a protection device for protecting an image capturer according to a second exemplary embodiment of the present disclosure
  • FIG. 7 shows a schematic diagram of an image captured by an image capturer according to a second exemplary embodiment of the present disclosure
  • FIG. 8 shows a flowchart of a step of detecting the position of the tip of a blade according to an exemplary embodiment of the present disclosure
  • FIG. 9 shows a flowchart of the step of identifying the edge of the tower according to an exemplary embodiment of the present disclosure
  • FIG. 10 shows a block diagram of a device for determining the tower headroom of a wind turbine according to an exemplary embodiment of the present disclosure
  • FIG. 11 shows a block diagram of a tower edge recognition module according to an exemplary embodiment of the present disclosure.
  • FIG. 1 shows a block diagram of a tower clearance monitoring system according to an exemplary embodiment of the present disclosure.
  • FIG. 2 shows a flowchart of a method for determining a tower headroom of a wind turbine according to an exemplary embodiment of the present disclosure.
  • the tower clearance monitoring system includes an image capturer 100 and a processor 200.
  • the image capturer 100 is used to capture images of the blades of the wind turbine during operation.
  • the processor 200 is configured to perform the method shown in FIG. 2 for determining the tower clearance of the wind turbine.
  • step S10 an image of the wind generating set during operation is acquired.
  • the acquired images include the tips of the blades of the wind turbine and the tower.
  • the image capturer 100 captures images of the blades of the wind turbine during operation, and the images captured by the image capturer 100 include images of the tips of the blades of the wind turbine and the tower tube for tower clearance analysis. Image, and then the blade tip and tower are identified for the image used for tower clearance analysis.
  • the image used for the tower clearance analysis is an image in which the tip of the blade of the wind turbine and the tower tube are included in the captured image.
  • the tower clearance refers to the distance between the tip of the blade and the surface of the tower during the rotation of the impeller, in order to determine the value of the tower clearance, it is necessary to pass the image of the tip of the blade and the tower containing the wind turbine Only through analysis can the tower clearance be determined.
  • the image capturer 100 may include, but is not limited to, a camera or a laser 2D (two-dimensional) scanner for capturing images of the blades of the wind turbine during operation.
  • the video camera can record the video of the blades of the wind turbine during the operation, and then identify each frame of the captured video, from each frame of the image Identify images used for tower clearance analysis.
  • images of the blades of a continuous multi-frame wind turbine during operation can be obtained by shooting video, and then for each frame of image, the image including the tip of the blade and the tower barrel can be identified for performing the subsequent tower Overhead clearance analysis. In this way, real-time monitoring of the tower headroom through video means is realized.
  • image recognition methods can be used to recognize the images captured by the image capturer 100 to identify the images including the tips of the blades of the wind turbine and the tower from the captured images, and to recognize Is determined as the image used for the tower clearance analysis.
  • the captured image may be identified through template matching. For example, multiple template images marked with the tips of the blades of the wind turbine and the tower can be established in advance, and the captured images are compared with the multiple template images, respectively.
  • multiple template images marked with the tip of the blade and the tower can be stacked on the image captured by the image capturer 100 for template matching, when any one of the multiple template images exists in the captured image
  • the matched image is determined as the image including the tip of the blade of the wind turbine and the tower, that is, the matched image is determined as the image used for the tower clearance analysis.
  • the above method of performing image recognition through template matching is only an example, and the present disclosure is not limited thereto, and other image recognition methods are also feasible.
  • the installation position of the image capturer 100 can be reasonably set to enable the image capturer 100 to capture images including the tips of the blades of the wind turbine and the tower. Two installation examples of the image capturer 100 are described below.
  • the image capturer 100 may be installed at the bottom of the nacelle of the wind turbine to capture images including the tips of the blades of the wind turbine and the tower.
  • FIG. 3 shows a schematic diagram of the installation position of the image capturer according to the first exemplary embodiment of the present disclosure.
  • the image capture device 100 may be provided at the bottom of the nacelle 3 of the wind turbine, that is, the image capture device 100 may be provided at the bottom of the nacelle 3 shell between the tower 1 and the hub, to When the blade 2 is rotated to an angle range that effectively measures the clearance, an image including the tip of the blade 2 and the tower 1 is captured.
  • the above-mentioned effective angular range for measuring the headroom may be a predetermined angular range.
  • the angle range of the effective measurement clearance may refer to a predetermined angle range near the azimuth angle of the impeller when the tip of the blade is perpendicular to the ground, in other words, it refers to the tower as the line of symmetry and radius and the center angle is predetermined Angled fan.
  • a bracket may be provided at the bottom of the nacelle 3 of the wind turbine to fix the image capturer 100 on the bracket.
  • the present disclosure is not limited to this, and the image capturer 100 may be directly installed on the bottom of the casing of the nacelle 3 without providing a bracket.
  • a focal length of more than 20 millimeters (mm) may be selected Camera.
  • the tip of the blade 2 and the tower 1 can be within the shooting range of the camera by adjusting the installation position of the camera and/or selecting a camera with a proper focal length, so as to capture a high-quality headroom for the tower Analyze the image.
  • the blade tip speed exceeds 80 seconds/meter (m/s) when the wind turbine is in full power. Therefore, the tip of the blade will appear in the shooting range of the camera for about 300 milliseconds (ms). You can capture images containing the tip of the blade, and you can select a camera with a frame rate of more than 20Hz.
  • the camera should also have a night vision function.
  • the camera's infrared fill light irradiation distance should reach 200 meters.
  • FIG. 4 shows a schematic diagram of an image captured by the image capturer 100 according to the first exemplary embodiment of the present disclosure.
  • FIG. 4 shows that when the image capturer 100 is set at the bottom of the nacelle 3, the image captured by the image capturer 100 including the tip A of the blade 2 and the tower 1 can be identified by recognizing the image shown in FIG. 4
  • the tower headroom S is determined, and the detailed process of tower headroom analysis for the image shown in FIG. 4 will be described later.
  • the image capturer 100 may be disposed in a designated area on the side of the wind turbine and at a predetermined distance from the wind turbine to capture the tip and tower of the blade that contains the wind turbine Image.
  • FIG. 5 shows a schematic diagram of an installation position of an image capturer according to a second exemplary embodiment of the present disclosure.
  • the image capturer 100 may be provided in a designated area on the side of the wind turbine, preferably, a bracket may be provided in the specified area to adjust the height and capture angle of the image capturer 100 to enable image capture
  • the imager 100 can capture images including the tip of the blade 2 and the tower 1.
  • a protection device may be provided in a designated area on the side of the wind turbine to reduce image capture in a harsh environment 100 captures the impact of the image process.
  • FIG. 6 shows a schematic diagram of a protection device for protecting an image capturer according to a second exemplary embodiment of the present disclosure.
  • the protection device may include a support plate 11 and a baffle 22.
  • the support plate 11 is used to fix the image collector 100, and the shooting height of the image capturer 100 can be adjusted by adjusting the height of the support plate 11 from the ground.
  • the baffle 22 is a three-sided baffle used to protect the image capturer 100 on three sides, so that the image capturer 100 is not easily affected by severe windy weather.
  • the baffle 22 may be a trapezoidal three-sided baffle.
  • protection device shown in FIG. 6 is only an example, and the present disclosure is not limited to this, and those skilled in the art may determine the shape and size of the support plate and the baffle plate according to needs.
  • those skilled in the art can also select other types of protection devices to protect the image capture device 100.
  • a shielding plate can be added above the image capture device 100, or a transparent protective cover can be provided around the image capture device 100. Wait.
  • FIG. 7 shows a schematic diagram of an image captured by an image capturer according to a second exemplary embodiment of the present disclosure.
  • FIG. 7 shows that when the image capturer 100 is disposed in a designated area on the side of the wind turbine, the image captured by the image capturer 100 includes the tip A of the blade and the tower tube.
  • the image recognition can determine the tower headroom S. The detailed process of tower headroom analysis for the image shown in FIG. 7 will be introduced later.
  • the image capturer is provided on the tower, so that the image capturer can be shot from top to bottom in the tower direction or from bottom to top in the tower direction to obtain an image that can be used for tower clearance analysis.
  • the processor 200 may be provided in the nacelle of the wind turbine generator for processing the image captured by the image capturer 100.
  • the processor 200 may also be located in the monitoring center (or dispatch center) of the wind farm.
  • the image capturer 100 may directly send the captured image to the processor 200, or the image capturer 100 may also The captured image is sent to the controller of the wind turbine, and the controller transmits the received image to the processor 200 for tower clearance analysis.
  • data transmission can be performed between the image capturer 100 and the processor 200 in a wired manner.
  • the image capturer 100 can send the captured image to the processor 200 in a bus manner.
  • the present disclosure is not limited to this, and data transmission between the image capturer 100 and the processor 200 may also be performed in a wireless manner.
  • the relative positional relationship between the image capturer 100 and the wind turbine is also fixed, so which area in the image captured by the image capturer 100 may be With the tower, which area may contain blades is also relatively fixed.
  • the method for determining the tower headroom of a wind turbine may further include: extracting the blade detection head from the acquired image for tower headroom analysis
  • the first sensitive area at the tip and the second sensitive area for identifying the edge of the tower can be subjected to tower clearance analysis for the extracted first and second sensitive areas.
  • step S20 the position of the tip of the blade of the wind turbine is determined from the acquired image.
  • the position of the tip of the blade of the wind turbine can be detected from the first sensitive area.
  • the blade tip feature points can be detected from the image used for tower clearance analysis (or in the first sensitive area), and the coordinates corresponding to the detected blade tip feature points can be used as the position of the blade tip.
  • various methods can be used to detect leaf tip feature points from the image, which is not limited in this disclosure.
  • those skilled in the art may also adopt other methods to detect the position of the tip of the blade from the first sensitive area.
  • the leaf tip feature point may be a pixel point in the image that satisfies any of the following cases: the pixel point with the largest gray gradient value in the image, the intersection of any two or more non-parallel straight lines, and the image Pixels where the gradient value of the gray scale is greater than the first set value and the change rate of the gradient direction is greater than the second set value.
  • a predetermined window may be used to traverse the first sensitive area to detect the position of the tip of the blade of the wind turbine from the first sensitive area.
  • traversing the first sensitive area refers to moving the predetermined window along a preset search route to implement leaf tip feature point detection on the entire first sensitive area.
  • the window size of the predetermined window may be set according to actual accuracy requirements.
  • the images contained in the predetermined window before and after the movement may not overlap at all, or may partially overlap, which is not limited in this disclosure, and those skilled in the art can select according to actual needs. That is to say, those skilled in the art can determine the size of the predetermined window and the size of the sliding displacement according to requirements, which is not limited in this disclosure.
  • step S201 using the current position of the predetermined window as a starting point, the predetermined window is slid in any direction on the first sensitive area.
  • step S202 for the sliding in each direction, the degree of gradation change of the pixels in the predetermined window before and after sliding is determined.
  • the degree of gradation change of the pixel point may refer to the gradient change speed of the gradation of the pixel point.
  • (u, v) represents the offset along the first predetermined direction and the second predetermined direction when the predetermined window slides
  • (x, y) represents the coordinate position of the corresponding pixel in the predetermined window
  • w (x, y) is the window function.
  • the window function can be set to a binary normal distribution with the center of the predetermined window as the origin.
  • I(x, y) represents the brightness (intensity) of the pixel
  • I(x+u, y+v) represents the brightness of the pixel after sliding (u, v) offset.
  • M is a 2 ⁇ 2 matrix
  • the expression of matrix M is:
  • the formula (2) can be used to determine the degree of gradation change of the pixels in the predetermined window before and after sliding.
  • the present disclosure is not limited to this, and those skilled in the art may also use other methods to determine the degree of gradation change.
  • step S203 it is judged whether the degree of gradation change satisfies the set condition when sliding in all directions.
  • the degree of gradation change of the pixels in the predetermined window before and after sliding meets the set condition means that the degree of gradation change corresponding to sliding in each direction is greater than the set change value (for example, sliding in each direction (The gradient speed of the gray scale of the corresponding pixel point is greater than the set change value).
  • step S204 is performed: changing the position of the predetermined window on the first sensitive area, and returning to step S201.
  • step S205 it is determined that there is a leaf tip feature in the image contained in the predetermined window at the current position point.
  • step S206 the leaf tip feature points are detected from the image included in the current position of the predetermined window, and the coordinates corresponding to the detected leaf tip feature points are determined as the position of the tip of the blade.
  • the blade tip feature points can be detected in the manner shown in steps S201 to S205 described above, and the coordinates corresponding to the detected blade tip feature points can be used as the position of the blade tip.
  • the position A of the tip of the blade can be detected from the first sensitive area in the above manner.
  • the step of detecting the leaf tip feature point from the image included in the current position of the predetermined window may include: maximizing the gradient value of the grayscale and/or the rate of change of the gradient direction in the image included in the image included in the predetermined window at the current position Of pixels are determined as leaf tip feature points.
  • blade tip feature point response function (blade tip feature point measurement function) R can be defined according to the following formula:
  • ⁇ 1 is the degree of gradient change in the first predetermined direction
  • ⁇ 2 is the degree of gradient change in the second predetermined direction
  • h is the response coefficient
  • the value of the leaf tip feature point response function R can be compared with a predetermined threshold, and when the value of R is greater than or equal to the predetermined threshold, the pixel corresponding to the local maximum of R is determined as the leaf tip Feature points.
  • detecting the position of the tip of the blade is to determine the degree of gray change of each pixel in the first sensitive area, preferably, it may be based on the gray change of each sub-pixel in the first sensitive area Degree to detect the position of the tip of the blade, which can improve the accuracy of detecting the position of the tip of the blade.
  • a sub-pixel is a pixel between two physical pixels (that is, the pixel mentioned above), and the sub-pixel exists within a gap from the physical pixel. That is to say, it is possible to perform leaf tip feature point detection based on sub-pixel points in the first sensitive area.
  • the step of detecting the leaf tip feature points from the first sensitive area may include: detecting a plurality of candidate leaf tip feature points from the first sensitive area; To determine the final leaf tip feature point.
  • the tip of the blade is close to the ground, and the tip of the blade is located at the bottom of the blade, so when multiple candidate blade tip feature points are detected, the candidate leaf closest to the ground The sharp feature point is most likely the tip of the blade.
  • the candidate closest to the ground can be selected from a plurality of candidate blade tip feature points based on the relative positional relationship between the image capturer 100 used to capture the image for tower clearance analysis and the wind turbine
  • the leaf tip feature point serves as the final leaf tip feature point (ie, the tip of the blade).
  • the point with the largest Y-axis coordinate value among the plurality of candidate blade tip feature points is determined as the final Characteristic point of the leaf tip.
  • the point with the smallest Y-axis coordinate value among the plurality of candidate leaf tip feature points (that is, the lowest candidate leaf tip feature point in the first sensitive area) is determined It is the final leaf tip feature point.
  • points A, A1, and A2 shown in FIG. 7 represent a plurality of points obtained when performing leaf tip feature point detection based on sub-pixel points in the first sensitive area.
  • the point A with the smallest Y-axis coordinate value may be determined as the final leaf tip feature point, that is, the coordinates corresponding to point A may be determined as the position of the tip of the blade.
  • FIG. 8 illustrates the steps of detecting the position of the tip of the blade from the first sensitive area by taking the example of extracting the first sensitive area from the image used for the tower clearance analysis.
  • the present disclosure is not limited to this, and the method of detecting the position of the tip of the blade shown in FIG. 8 is also applicable to the case of detecting the position of the tip of the blade from the image used for tower clearance analysis (from the captured image). In this case, it is necessary to use the predetermined window to traverse the entire image for tower clearance analysis. When the predetermined window is at any position on the image for tower clearance analysis, you can use the method shown in Figure 8 to The position of the tip of the blade is detected in the image contained in the window.
  • a straight line can be identified from an image through a straight line detection method to determine the intersection of two or more straight lines as a leaf tip feature point.
  • step S30 the edge of the tower is identified from the acquired image.
  • the designated point in the image can be used as the edge of the tower.
  • the designated point may be a pixel point in the image corresponding to a point on the tower used for determining the tower clearance determined on the basis of the relative relationship between the image capturer and the tower of the wind turbine. That is to say, the designated point may be a pixel point corresponding to the image in the position where the blade is most likely to contact the tower during operation.
  • point B is a designated point that is the edge of the tower in the image.
  • the edge of the tower can be identified by performing edge detection (or straight line detection) on the image.
  • the edge of the tower can be identified from the second sensitive area.
  • various straight line detection methods may be used to detect a straight line from the second sensitive area, and use the detected straight line as the edge of the tower, but the present disclosure is not limited to this, and other methods may be used to identify from the second sensitive area
  • the edge of the tower is determined by, for example, identifying the preset mark for indicating the edge of the tower from the second sensitive area.
  • FIG. 9 shows a flowchart of the step of identifying the edge of the tower according to an exemplary embodiment of the present disclosure.
  • step S301 multiple edge feature points are extracted from the second sensitive area.
  • the image corresponding to the second sensitive area may be converted into a grayscale image, and edge feature points may be extracted from the converted grayscale image.
  • edge feature points can also be extracted in other ways.
  • step S302 the extracted multiple edge feature points are mapped into the parameter space, and corresponding straight lines are drawn in the parameter space based on the multiple edge feature points.
  • these three straight lines intersect at the same point (1,0)
  • the horizontal and vertical coordinates of this point are the slope and intercept of the straight line in the image space. That is to say, when the intersection point of multiple straight lines is found in the parameter space, the straight line in the image space can be determined.
  • step S303 at least one aggregation point in the parameter space is determined.
  • the at least one gathering point is a point where a straight line exceeds a predetermined number.
  • step S304 at least one straight line in the second sensitive area is determined according to the coordinate value of the at least one gathering point, and the edge of the tower barrel is determined based on the at least one straight line.
  • the step of determining at least one straight line in the second sensitive area according to the coordinate value of at least one gathering point may include: for each gathering point, use the abscissa of the gathering point as the slope of the straight line, and the longitudinal of the gathering point The coordinates are used as the intercept of the straight line to obtain a straight line corresponding to the gathering point in the second sensitive area. That is to say, the edge feature points can be extracted from the second sensitive area (ie, the image space) based on the predetermined coordinate system of the image space, and after determining the aggregation point, the corresponding straight line can be drawn under the predetermined coordinate system.
  • the step of determining the edge of the tower barrel based on at least one straight line may include: determining the edge of the tower barrel based on the fitted straight line obtained after the fitting by fitting the determined at least one straight line.
  • a straight line with a relative distance less than a specified distance can be selected from at least one straight line for fitting.
  • the obtained fitted straight line is used as the edge of the tower, and two pieces are obtained after fitting.
  • fitting a straight line use the line connecting the midpoints of the two fitted straight lines as the edge of the tower.
  • the line connecting the midpoints of the two straight lines may be determined as the edge of the tower.
  • a straight line L in the second sensitive area is obtained by fitting, and the straight line L is determined as the edge of the tower barrel.
  • the distance from the point to the straight line can be used to determine the distance from the tip of the blade to the edge of the tower.
  • the above manner of determining the tower edge by fitting at least one straight line is only an example, and the present disclosure is not limited thereto, and other ways of determining the tower edge are also feasible.
  • the length of each straight line can be calculated, and the longest straight line is selected as the edge of the tower.
  • FIG. 9 takes the example of extracting the second sensitive area from the image used for the tower headroom analysis as an example to introduce the steps of identifying the edge of the tower barrel from the second sensitive area.
  • the present disclosure is not limited to this, and the method of identifying the edge of the tower shown in FIG. 9 is also applicable to the case of identifying the edge of the tower from the image used for the tower clearance analysis.
  • multiple edge feature points can be extracted from the image used for tower headroom analysis to identify the edge of the tower.
  • edge information can be extracted from the second sensitive area, such as extracting multiple edge points, and fitting the multiple edge points to obtain an edge straight line as the edge of the tower.
  • step S40 according to the determined position of the tip of the blade and the identified cylinder edge of the tower, the distance from the blade tip to the edge of the tower cylinder is calculated to obtain the tower headroom.
  • the vertical distance from the position of the tip of the blade (such as the coordinate position) to the straight line corresponding to the edge of the tower can be calculated as the tower clearance.
  • the step of calculating the distance from the tip of the blade to the edge of the tower to obtain tower clearance may include: according to the determined position and identification of the tip of the blade Calculate the pixel distance from the tip of the blade to the edge of the tower; based on the correspondence between the predetermined pixel distance between any two pixels and the actual distance, use the tip of the blade to the edge of the tower The pixel distance calculates the actual distance from the tip of the blade to the edge of the tower, and determines the actual distance as the tower clearance.
  • any two pixels may be two adjacent pixels or two pixels specified on the image.
  • FIG. 10 shows a block diagram of an apparatus for determining a tower headroom of a wind turbine according to an exemplary embodiment of the present disclosure.
  • the apparatus for determining the tower headroom of a wind turbine includes an image acquisition module 10, a blade tip detection module 20, a tower edge recognition module 30 and a tower headroom determination module 40.
  • the image acquisition module 10 acquires images of the wind turbine during operation.
  • the acquired images include the tips of the blades of the wind turbine and the tower.
  • the image acquisition module 10 may acquire images of the blades of the wind turbine during operation, and determine that the images of the blades during the operation include the tips of the blades of the wind turbine and the tower tube for tower clearance analysis Image.
  • an image capturer can be used to capture images of the blades of the wind turbine during operation, and then the image capturer sends the captured images to the image acquisition module 10.
  • the installation position of the image capturer can be reasonably set so that the image capturer can capture images including the tips of the blades of the wind turbine and the tower.
  • the following describes two installation examples of the image capturer.
  • the image capturer may be arranged at the bottom of the nacelle of the wind turbine to capture images including the tips of the blades of the wind turbine and the tower.
  • the image capture device can be located on the side of the wind turbine and the distance between the wind turbine and the wind turbine is a predetermined distance in a specified area to capture the tip of the wind turbine blade and tower image.
  • the device for determining the tower headroom of a wind turbine may further include: a sensitive area extraction module (not shown in the figure), which is used to detect blades from the acquired image The first sensitive area at the tip of the and the second sensitive area used to identify the edge of the tower, the tower headroom analysis can be performed on the extracted first and second sensitive areas later.
  • a sensitive area extraction module (not shown in the figure), which is used to detect blades from the acquired image The first sensitive area at the tip of the and the second sensitive area used to identify the edge of the tower, the tower headroom analysis can be performed on the extracted first and second sensitive areas later.
  • the blade tip detection module 20 determines the position of the tip of the blade of the wind turbine from the acquired image.
  • the blade tip detection module 20 may determine the position of the tip of the blade of the wind turbine from the first sensitive area.
  • the blade tip detection module 20 can use a predetermined window to traverse the first sensitive area.
  • the predetermined window is at any position on the first sensitive area, the blade tip detection module 20 can be included from the predetermined window in the following manner The position of the tip of the blade is detected in the image:
  • the blade tip detection module 20 may determine the pixel point with the largest gradient value of grayscale and/or the highest rate of change in gradient direction in the image contained when the predetermined window is at any position as the blade tip feature point.
  • the present disclosure is not limited to this, and the intersection point of any two or more non-parallel straight lines may also be determined as the blade tip characteristic point.
  • the blade tip detection module 20 may detect multiple candidate leaf tip feature points from the first sensitive area, and determine the final leaf tip feature point from the multiple candidate leaf tip feature points according to a preset condition.
  • the blade tip detection module 20 selects the candidate blade tip feature closest to the ground from a plurality of candidate blade tip feature points based on the relative positional relationship between the image capturer used to capture the image for tower clearance analysis and the wind turbine The point is used as the final leaf tip characteristic point.
  • the blade tip detection module 20 may determine the point with the largest Y-axis coordinate value among the plurality of candidate blade tip feature points as the final blade tip feature point.
  • the blade tip detection module 20 may determine the point with the smallest Y-axis coordinate value among the plurality of candidate blade tip feature points as the final blade tip feature point.
  • the tower edge recognition module 30 recognizes the edge of the tower from the acquired image.
  • the tower edge identification module 30 may identify the edge of the tower from the second sensitive area.
  • the tower edge recognition module 30 may use the designated point in the image as the edge of the tower.
  • the designated point may be a pixel point in the image corresponding to a point on the tower used for determining the tower clearance determined on the basis of the relative relationship between the image capturer and the tower of the wind turbine. That is to say, the designated point may be a pixel point corresponding to the image in the position where the blade is most likely to contact the tower during operation.
  • the tower edge recognition module 30 may identify the edge of the tower by performing edge detection on the image.
  • the tower edge identification module 30 may use various image recognition methods to identify the edge of the tower from the second sensitive area. The function of the tower edge recognition module 30 will be described below with reference to FIG. 11. It should be understood that the method for identifying the edge of the tower shown in FIG. 11 is only a preferred example, and other image recognition methods for identifying the edge of the tower are also feasible.
  • FIG. 11 shows a block diagram of a tower edge recognition module 30 according to an exemplary embodiment of the present disclosure.
  • the tower edge recognition module 30 may include a feature extraction submodule 301, a conversion submodule 302, an aggregation point determination submodule 303 and an edge determination submodule 304.
  • the feature extraction submodule 301 extracts multiple edge feature points from the second sensitive area.
  • the feature extraction submodule 301 may convert the image corresponding to the second sensitive area into a grayscale image, and extract edge feature points from the converted grayscale image.
  • the conversion sub-module 302 maps the extracted multiple edge feature points into the parameter space, and draws corresponding straight lines in the parameter space based on the multiple edge feature points.
  • the aggregation point determination submodule 303 determines at least one aggregation point in the parameter space.
  • the at least one gathering point is a point where a straight line exceeds a predetermined number.
  • step 303 of FIG. 9 Since the detailed process of the aggregation point determination sub-module 303 determining at least one aggregation point has been described in step 303 of FIG. 9, the content of this part of the present disclosure will not be repeated here.
  • the edge determination submodule 304 determines at least one straight line in the second sensitive area according to the coordinate value of the at least one gathering point, and determines the edge of the tower barrel based on the at least one straight line.
  • the edge determination submodule 304 may use the horizontal coordinate of the aggregation point as the slope of the straight line, and the vertical coordinate of the aggregation point as the intercept of the straight line to obtain the aggregation point in the second sensitive area. Corresponding straight line.
  • the edge determination submodule 304 may select a straight line with a relative distance less than a specified distance from at least one straight line to perform fitting.
  • a fitted straight line is obtained after fitting, the obtained fitted straight line is used as the edge of the tower.
  • two fitted straight lines are obtained after closing, the line connecting the midpoints of the two fitted straight lines is taken as the edge of the tower.
  • the tower clearance determination module 40 calculates the distance from the tip of the blade to the edge of the tower based on the determined position of the tip of the blade and the identified edge of the tower to obtain the tower clearance. For example, the tower headroom determination module 40 may calculate the vertical distance from the position of the tip of the blade to the straight line corresponding to the edge of the tower barrel as the tower headroom.
  • the tower clearance determination module 40 may calculate the pixel distance from the tip of the blade to the edge of the tower according to the determined position of the blade tip and the identified edge of the tower, based on the predetermined pixel distance between any two pixels and The corresponding relationship of the actual distance is used to calculate the actual distance from the tip of the blade to the edge of the tower using the pixel distance from the tip of the blade to the edge of the tower, and the actual distance is determined as the tower headroom.
  • Exemplary embodiments according to the present disclosure also provide a computer-readable storage medium storing a computer program.
  • the computer-readable storage medium stores a computer program that, when executed by the processor, causes the processor to execute the above method for determining the tower headroom of the wind turbine.
  • the computer-readable recording medium is any data storage device that can store data read by a computer system. Examples of computer-readable recording media include read-only memory, random-access memory, read-only optical disks, magnetic tapes, floppy disks, optical data storage devices, and carrier waves (such as data transmission through the Internet via wired or wireless transmission paths).
  • the tower headroom of the wind power generator can be monitored in real time to effectively avoid the losses caused by the blade sweeping tower.
  • the method and device for determining the tower headroom of the wind power generator set according to the exemplary embodiments of the present disclosure can not only fully realize the tower by properly designing the bracket for supporting the image capture device and selecting the installation position of the image capture device
  • the function of headroom video monitoring can also ensure the safe operation of wind turbines.
  • the tower headroom of the wind power generator can be obtained relatively simply, without manual measurement, and is convenient and quick.
  • the tip of the blade is detected based on the blade tip feature point detection method, and the edge of the tower barrel can also be identified by a straight line detection method, thereby The distance between the tip of the blade and the edge of the tower.

Abstract

一种确定风力发电机组的塔架净空的方法和装置，该方法包括：获取风力发电机组在运行过程中的图像（S10），图像包括风力发电机组的叶片（2）的尖端以及塔筒（1）；从获取的图像中确定风力发电机组的叶片（2）的尖端的位置（S20）；从获取的图像中识别塔筒（1）的边缘（S30）；根据确定的叶片（2）的尖端的位置和识别的塔筒（1）的边缘，计算叶片（2）的尖端到塔筒（1）的边缘的距离以获得塔架净空（S40）。采用该确定风力发电机组的塔架净空的方法和装置，能够实时确定风力发电机组的塔架净空，以有效避免叶片（2）扫塔情况的发生。

Description

确定风力发电机组的塔架净空的方法和装置 技术领域
本公开总体说来涉及风电技术领域，更具体地讲，涉及一种确定风力发电机组的塔架净空的方法和装置。
背景技术
风力发电机组的塔架净空是指叶轮在旋转过程中叶片的尖端到塔筒表面的距离。对于风力发电机组而言，如果一旦发生叶片扫塔，则需要更换叶片，而单只叶片的成本较高，这会增加维修成本。同时在更换叶片期间需机组停机，而机组停机期间还会导致发电量的损失，因此一旦发生叶片扫塔会为风电场带来较大的经济损失。
而目前风力发电机组的塔架净空无法通过测量工具测量，导致无法实时地获取风力发电机组的塔架净空。
发明内容
本公开的示例性实施例提供一种确定风力发电机组的塔架净空的方法和装置，以解决现有技术中无法对风力发电机组的塔架净空进行测量的技术问题。
在一个总体方面，提供一种确定风力发电机组的塔架净空的方法，所述方法包括：获取风力发电机组在运行过程中的图像，所述图像包括所述风力发电机组的叶片的尖端以及塔筒；从获取的图像中确定风力发电机组的叶片的尖端的位置；从获取的图像中识别塔筒的边缘；根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
在另一总体方面，提供一种确定风力发电机组的塔架净空的装置，所述装置包括：图像获取模块，获取风力发电机组在运行过程中的图像，所述图像包括所述风力发电机组的叶片的尖端以及塔筒；叶尖检测模块，从获取的图像中确定风力发电机组的叶片的尖端的位置；塔筒边缘识别模块，从获取 的图像中识别塔筒的边缘；塔架净空确定模块，根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
在另一总体方面，提供一种塔架净空监测系统，所述塔架净空监测系统包括：图像捕获器，用于捕获风力发电机组的叶片在运行过程中的图像；处理器，被配置为：从所捕获的图像中获取包括所述风力发电机组的叶片的尖端以及塔筒的图像；从获取的图像中确定风力发电机组的叶片的尖端的位置；从获取的图像中识别塔筒的边缘；根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
在另一总体方面，提供一种存储有计算机程序的计算机可读存储介质，当所述计算机程序在被处理器执行时实现上述的确定风力发电机组的塔架净空的方法。
采用本公开示例性实施例的确定风力发电机组的塔架净空的方法和装置，能够实时确定风力发电机组的塔架净空，以有效避免叶片扫塔情况的发生。
附图说明
通过下面结合附图进行的描述，本公开的上述和其他目的和特点将会变得更加清楚，其中：
图1示出根据本公开示例性实施例的塔架净空监测系统的框图；
图2示出根据本公开示例性实施例的确定风力发电机组的塔架净空的方法的流程图；
图3示出根据本公开第一示例性实施例的图像捕获器安装位置的示意图；
图4示出根据本公开第一示例性实施例的图像捕获器所捕获的图像的示意图；
图5示出根据本公开第二示例性实施例的图像捕获器安装位置的示意图；
图6示出根据本公开第二示例性实施例的用于保护图像捕获器的保护装置的示意图；
图7示出根据本公开第二示例性实施例的图像捕获器所捕获的图像的示意图；
图8示出根据本公开示例性实施例的检测叶片的尖端的位置的步骤的流程图；
图9示出根据本公开示例性实施例的识别塔筒的边缘的步骤的流程图；
图10示出根据本公开示例性实施例的确定风力发电机组的塔架净空的装置的框图；
图11示出根据本公开示例性实施例的塔筒边缘识别模块的框图。
具体实施方式
现在，将参照附图更充分地描述不同的示例实施例，一些示例性实施例在附图中示出。
图1示出根据本公开示例性实施例的塔架净空监测系统的框图。图2示出根据本公开示例性实施例的确定风力发电机组的塔架净空的方法的流程图。
如图1所示，根据本公开示例性实施例的塔架净空监测系统包括图像捕获器100和处理器200。图像捕获器100用于捕获风力发电机组的叶片在运行过程中的图像，处理器200被配置为执行图2所示的确定风力发电机组的塔架净空的方法。
下面结合图1和图2来介绍确定风力发电机组的塔架净空的过程。
参照图2，在步骤S10中，获取风力发电机组在运行过程中的图像。获取的图像中包括风力发电机组的叶片的尖端以及塔筒。
例如，图像捕获器100捕获风力发电机组的叶片在运行过程中的图像，将图像捕获器100所捕获的图像中包含风力发电机组的叶片的尖端和塔筒的图像确定为用于塔架净空分析的图像，后续针对用于塔架净空分析的图像进行叶片的尖端和塔筒的识别。
也就是说，用于塔架净空分析的图像为所拍摄的图像中包含风力发电机组的叶片的尖端和塔筒的图像。这里，由于塔架净空是指叶轮在旋转过程中叶片的尖端到塔筒表面的距离，因此，为确定出塔架净空的值，需通过对包含风力发电机组的叶片的尖端和塔筒的图像进行分析才能确定出塔架净空。
作为示例，图像捕获器100可包括但不限于摄像机或者激光2D(二维)扫描仪，以用于捕获风力发电机组的叶片在运行过程中的图像。
在一优选实施例中，当图像捕获器100为摄像机时，摄像机可以拍摄风力发电机组的叶片运行过程中的视频，后续针对拍摄的视频中的每一帧图像进行识别，从每一帧图像中识别出用于塔架净空分析的图像。也就是说，可以通过拍摄视频的方式获得连续多帧风力发电机组的叶片在运行过程中的图 像，然后针对每一帧图像识别出包括叶片的尖端以及塔筒的图像以用于执行后续的塔架净空分析。这样，实现了通过视频手段对塔架净空的实时监测。
应理解，可利用各种图像识别方式来对图像捕获器100所捕获的图像进行识别，以从所捕获的图像中识别出包含风力发电机组的叶片的尖端和塔筒的图像，并将识别出的图像确定为用于塔架净空分析的图像。
作为示例，可通过模板匹配的方式来对所捕获的图像进行识别。例如，可预先建立标记了风力发电机组的叶片的尖端和塔筒的多个模板图像，将所捕获的图像分别与多个模板图像进行比对。
具体地，可将标记了叶片的尖端和塔筒的多个模板图像分别叠放在图像捕获器100所捕获的图像上进行模板匹配，当所捕获的图像中存在与多个模板图像中的任意一个模板图像匹配的图像时，将该匹配的图像确定为包含风力发电机组的叶片的尖端和塔筒的图像，即，将匹配的图像确定为用于塔架净空分析的图像。但应理解，上述通过模板匹配的方式进行图像识别的方式仅为一示例，本公开不限于此，其他图像识别方法也是可行的。
可以通过合理的设置图像捕获器100的安装位置，来使得图像捕获器100能够捕获到包含风力发电机组的叶片的尖端和塔筒的图像。下面介绍图像捕获器100的两个安装示例。
第一种情况，图像捕获器100可被设置在风力发电机组的机舱底部，以捕获到包含风力发电机组的叶片的尖端和塔筒的图像。
图3示出根据本公开第一示例性实施例的图像捕获器安装位置的示意图。
如图3所示，可以在风力发电机组的机舱3底部设置图像捕获器100，也就是说，可以在机舱3外壳底部上处于塔筒1与轮毂之间的区域内设置图像捕获器100，以在叶片2旋转至有效测量净空的角度范围内时，捕获包含叶片2的尖端和塔筒1的图像。
也就是说，可以通过调整图像捕获器100与风力发电机组之间的相对位置关系，使得叶片位于有效测量净空的角度范围内时，叶片的尖端正好能够处于图像捕获器100的图像捕获范围内。
这里，上述有效测量净空的角度范围可以是预定的角度范围。例如，该有效测量净空的角度范围可指当叶片的尖端垂直于地面时所处的叶轮方位角附近的预定角度范围，换句话说，是指以塔架为对称线和半径且中心角为预定角度的扇形。
优选地，可在风力发电机组的机舱3底部设置一支架，将图像捕获器100固定于该支架上。但本公开不限于此，也可以不设置支架，直接将图像捕获器100安装在机舱3外壳底部。
这里，以图像捕获器100为摄像机为例，由于风力发电机组的叶片长度一般超过60米(m)，为了实现对于60米外塔架净空的精确测量，可选取焦距超过20毫米(mm)的摄像机。也就是说，可以通过调整摄像机的安装位置和/或选择合适焦距的摄像机来使得叶片2的尖端和塔筒1能够处在摄像机的拍摄范围内，以捕获到质量较高的用于塔架净空分析的图像。
由于风力发电机组在满发状态时叶尖速度超过80秒/米(m/s)，因此，叶片的尖端在摄像机的拍摄范围内会出现大约300毫秒(ms)，为了保证在300毫秒内摄像机可以捕获到包含叶片的尖端的图像，可选取帧率达到20Hz以上的摄像机。
由于在夜间也需要对塔架净空进行实时监测，因此优选地，摄像机还应具备夜视功能，作为示例，摄像机的红外补光灯照射距离应达到200米。
应理解，上述所示的在本公开示例性实施例中选取的摄像机的各项参数值和功能仅为示例，本公开不限于此，本领域技术人员可以根据实际需求来调整上述各参数值以选取适合的摄像机，只要使得摄像机能够捕获到包含风力发电机组的叶片的尖端和塔筒的图像即可。
图4示出根据本公开第一示例性实施例的图像捕获器100所捕获的图像的示意图。
图4示出的是当图像捕获器100被设置在机舱3底部时，图像捕获器100捕获到的包含叶片2的尖端A和塔筒1的图像，通过对图4所示的图像进行识别能够确定出塔架净空S，后续将介绍针对图4所示的图像进行塔架净空分析的详细过程。
第二种情况，图像捕获器100可被设置在位于风力发电机组侧面，且与风力发电机组之间的距离为预定距离的指定区域内，以捕获到包含风力发电机组的叶片的尖端和塔筒的图像。
图5示出根据本公开第二示例性实施例的图像捕获器安装位置的示意图。
如图5所示，可以在风力发电机组的侧面的指定区域内设置图像捕获器100，优选地，可以在指定区域内设置一支架来调整图像捕获器100的高度和捕获角度，以使图像捕获器100能够捕获到包含叶片2的尖端和塔筒1的图 像。
由于风力发电机组的运行环境比较恶劣，为使图像捕获器100能够捕获到清晰、稳定的图像，可以在风力发电机组的侧面的指定区域内设置一保护装置，用以减小恶劣环境对图像捕获器100捕获图像过程的影响。
图6示出根据本公开第二示例性实施例的用于保护图像捕获器的保护装置的示意图。
如图6所示，保护装置可包括支撑板11和挡板22，支撑板11用于固定图像采集器100，并可通过调整支撑板11距离地面的高度来调整图像捕获器100的拍摄高度。挡板22为三面挡板，用于对图像捕获器100进行三面保护，使得图像捕获器100不易受到恶劣大风天气的影响。作为示例，挡板22可为呈梯形的三面挡板。
应理解，图6所示的保护装置的形式仅为示例，本公开不限于此，本领域技术人员可以根据需要来决定支撑板和挡板的形状和大小尺寸。此外，本领域技术人员还可以选取其他样式的保护装置对图像捕获器100进行保护，例如，可在图像捕获器100的上方增加一遮挡板，或者在图像捕获器100的外围设置一透明保护罩等。
图7示出根据本公开第二示例性实施例的图像捕获器所拍摄的图像的示意图。
图7示出的是当图像捕获器100被设置在风力发电机组的侧面的指定区域内时，图像捕获器100捕获到的包含叶片的尖端A和塔筒的图像，通过对图7所示的图像进行识别能够确定出塔架净空S，后续将介绍针对图7所示的图像进行塔架净空分析的详细过程。
应理解，上述介绍的图像捕获器100的安装位置的两个示例仅为优选示例，本公开不限于此，本领域技术人员可以根据实际需要来改变图像捕获器的安装位置，例如，还可以将图像捕获器设置在塔筒上，使得图像捕获器可沿塔筒方向从上向下拍摄或者沿塔筒方向从下向上拍摄，以获得能够用于塔架净空分析的图像。
作为示例，处理器200可设置在风力发电机组的机舱内，以用于对图像捕获器100所捕获的图像进行处理。或者，处理器200还可设置在风电场的监控中心(或调度中心)，此时，图像捕获器100可将所捕获的图像直接发送至处理器200，或者，图像捕获器100也可以将所捕获的图像发送至风力发 电机组的控制器，由控制器将接收到的图像传送至处理器200以进行塔架净空分析。
此外，为了保证对风力发电机组的塔架净空监测的实时性，还需要尽可能的减少整体塔架净空监测系统在通讯传输时的耗时。优选地，图像捕获器100与处理器200之间可以通过有线方式进行数据传输，例如，图像捕获器100可以以总线方式将所捕获的图像发送给处理器200。但本公开不限于此，图像捕获器100与处理器200之间也可以采用无线方式进行数据传输。
这里，由于图像捕获器100的设置位置是固定的，那么图像捕获器100与风力发电机组之间的相对位置关系也是固定的，因此，在该图像捕获器100所捕获的图像中哪一区域可能含有塔筒，哪一区域可能含有叶片，也是相对固定。
基于此，在一优选实施例中，根据本公开示例性实施例的确定风力发电机组的塔架净空的方法可还包括：从获取的用于塔架净空分析的图像中提取用于检测叶片的尖端的第一敏感区域和用于识别塔筒的边缘的第二敏感区域，后续可以针对提取的第一敏感区域和第二敏感区域进行塔架净空分析。
返回图1，在步骤S20中，从获取的图像中确定风力发电机组的叶片的尖端的位置。
针对上述对用于塔架净空分析的图像进行第一敏感区域提取的情况，可从第一敏感区域中检测风力发电机组的叶片的尖端的位置。
在一优选实施例中，可从用于塔架净空分析的图像中(或者第一敏感区域中)检测叶尖特征点，将检测到的叶尖特征点对应的坐标作为叶片的尖端的位置。这里，可利用各种方法来从图像中检测出叶尖特征点，本公开对此不做限定。此外，本领域技术人员也可以采用其他方式来从第一敏感区域中检测叶片的尖端的位置。
这里，叶尖特征点可为图像中满足以下情况中的任一情况的像素点：图像中灰度的梯度值最大的像素点、任意两条或者两条以上不平行的直线的交点、图像中灰度的梯度值大于第一设定值且梯度方向的变化速率大于第二设定值的像素点。
下面参照图8来介绍从第一敏感区域中检测风力发电机组的叶片的尖端的位置的步骤。
图8示出根据本公开示例性实施例的检测叶片的尖端的位置的步骤的流 程图。在本公开示例性实施例中，可使用一预定窗口遍历第一敏感区域，以从第一敏感区域中检测出风力发电机组的叶片的尖端的位置。
这里，遍历第一敏感区域是指沿着预先设定好的搜索路线，移动该预定窗口，以实现对整个第一敏感区域进行叶尖特征点检测。
这里，可根据实际精度需求来设定该预定窗口的窗口尺寸。优选地，在移动该预定窗口时，移动前后该预定窗口所包含的图像可以完全不重叠，也可以部分重叠，本公开对此不做限定，本领域技术人员可以根据实际需求来进行选择。也就是说，本领域技术人员可以根据需求来确定预定窗口的尺寸和滑动位移的大小，本公开对此不做限定。
参照图8，在步骤S201中，以预定窗口所在的当前位置为起点，在第一敏感区域上沿任意方向滑动预定窗口。
在步骤S202中，针对沿每个方向的滑动，确定滑动前与滑动后预定窗口内的像素点的灰度变化程度。作为示例，像素点的灰度变化程度可指像素点的灰度的梯度变化速度。
例如，将预定窗口沿第一预定方向平移u，沿第二预定方向平移v，可产生灰度变化E(u,v)，如下式表示：
Figure PCTCN2019109391-appb-000001
公式(1)中，(u,v)表示预定窗口滑动时沿第一预定方向和第二预定方向的偏移量，(x,y)表示预定窗口内所对应的像素点的坐标位置，w(x,y)为窗口函数，作为示例，可将窗口函数设定为以预定窗口中心为原点的二元正态分布。I(x,y)表示像素点的亮度(强度)，I(x+u,y+v)表示滑动(u,v)偏移量后的像素点的亮度。
由于，
Figure PCTCN2019109391-appb-000002
因此，可得到如下形式的E(u,v)近似表达式：
Figure PCTCN2019109391-appb-000003
其中，M是一个2×2的矩阵，矩阵M的表达式为：
Figure PCTCN2019109391-appb-000004
由此，可利用公式(2)来确定滑动前与滑动后预定窗口内的像素点的灰度变化程度。但本公开不限于此，本领域技术人员也可以采用其他方式来确定灰度变化程度。
在步骤S203中，判断沿所有方向的滑动，灰度变化程度是否均满足设定条件。
滑动前与滑动后预定窗口内的像素点的灰度变化程度均满足设定条件是指沿每个方向的滑动对应的灰度变化程度均大于设定变化值(例如，沿每个方向的滑动对应的像素点的灰度的梯度变化速度均大于设定变化值)。
如果针对沿所有方向的滑动，滑动前与滑动后预定窗口内的像素点的灰度变化程度没有均满足设定条件，即，存在沿至少一个方向的滑动，滑动前与滑动后预定窗口内的像素点的灰度变化程度没有满足设定条件，则执行步骤S204：改变预定窗口在第一敏感区域上的位置，并返回执行步骤S201。
例如，当存在沿至少一个方向的滑动对应的灰度变化程度没有大于设定变化值时，确定滑动前与滑动后预定窗口内的像素点的灰度变化程度没有均满足设定条件。
如果针对沿所有方向的滑动，滑动前与滑动后预定窗口内的像素点的灰度变化程度均满足设定条件，则执行步骤S205：确定预定窗口在当前位置时包含的图像中存在叶尖特征点。
在步骤S206中，从预定窗口在当前位置时包含的图像中检测叶尖特征点，并将检测到的叶尖特征点对应的坐标确定为叶片的尖端的位置。
也就是说，可通过上述步骤S201～S205所示的方式进行叶尖特征点检测，将检测出的叶尖特征点对应的坐标作为叶片的尖端的位置。例如，以图4所示的图像为例，通过上述方式可从第一敏感区域中检测出叶片的尖端的位置A。
作为示例，从预定窗口在当前位置时包含的图像中检测叶尖特征点的步 骤可包括：将预定窗口在当前位置时包含的图像中灰度的梯度值最大和/或梯度方向的变化速率最高的像素点确定为叶尖特征点。
优选地，可按照如下公式来定义叶尖特征点响应函数(叶尖特征点度量函数)R：
R＝λ 1λ 2-h(λ 1+λ 2) 2                      (4)
公式(4)中，λ 1为在第一预定方向上的梯度变化程度，λ 2为在第二预定方向上的梯度变化程度，h为响应系数。
在一优选实施例中，可将叶尖特征点响应函数R的值与预定阈值进行比较，当R的值大于或等于预定阈值时，将R的局部极大值对应的像素点确定为叶尖特征点。
应理解，在上述检测叶片的尖端的位置的示例中是确定第一敏感区域中的各像素点的灰度变化程度，优选地，可基于第一敏感区域中的各亚像素点的灰度变化程度来检测叶片的尖端的位置，这样可以提高对叶片的尖端的位置的检测的精确性。
这里，亚像素点(Sub-Pixel)为两个物理像素(即，上述提及的像素点)之间的像素，亚像素点存在与物理像素的间隙内。也就是说，可以基于第一敏感区域中的亚像素点来进行叶尖特征点检测。
应理解，由于叶片涂装等问题，有时会将叶片的尖端部分涂装成红白颜色交替的图案，这会导致叶尖特征点检测存在误差，从而可能导致从第一敏感区域中检测出多个叶尖特征点。为了解决这个技术问题，从第一敏感区域中检测叶尖特征点的步骤可包括：从第一敏感区域中检测出多个候选叶尖特征点；根据预设条件从多个候选叶尖特征点中确定最终的叶尖特征点。
这里，由于在计算塔架净空时，叶片的尖端是接近于垂直指向地面的，而叶片的尖端位于叶片的最下端，因此当检测出多个候选叶尖特征点时，最靠近地面的候选叶尖特征点最可能是叶片的尖端。
在此情况下，可根据用于捕获用于塔架净空分析的图像的图像捕获器100与风力发电机组之间的相对位置关系，来从多个候选叶尖特征点中选择最靠近地面的候选叶尖特征点作为最终的叶尖特征点(即，叶片的尖端)。
针对图像捕获器100被设置在机舱底部的情况，将多个候选叶尖特征点中的Y轴坐标值最大的点(即，处于第一敏感区域中最上端的候选叶尖特征点)确定为最终的叶尖特征点。
针对图像捕获器100被设置在指定区域内的情况，将多个候选叶尖特征点中的Y轴坐标值最小的点(即，处于第一敏感区域中最下端的候选叶尖特征点)确定为最终的叶尖特征点。
在此情况下，以图7所示的图像为例，图7中所示的点A、A1、A2表示基于第一敏感区域中的亚像素点来进行叶尖特征点检测时获得的多个候选叶尖特征点，此时，可将Y轴坐标值最小的点A确定为最终的叶尖特征点，即，将点A对应的坐标确定为叶片的尖端的位置。
应理解，图8是以从用于塔架净空分析的图像中提取了第一敏感区域为例来介绍从第一敏感区域中检测叶片的尖端的位置的步骤。但本公开不限于此，图8所示的检测叶片的尖端的位置的方法也适用于从用于塔架净空分析的图像中(从所捕获的图像中)检测叶片的尖端的位置的情况。在此情况下，需使用预定窗口遍历整个用于塔架净空分析的图像，当预定窗口处于用于塔架净空分析的图像上的任一位置时，可通过图8所示的方法来从预定窗口内包含的图像中检测叶片的尖端的位置。
此外，图8所示的检测叶片的尖端的位置的方式仅为示例，本公开不限于此，本领域技术人员还可以采用其他图像识别方法来从图像中检测出叶片的尖端。例如，可以通过直线检测方式，从图像中识别出直线，以将两条或者两条以上直线的交点确定为叶尖特征点。
返回图1，在步骤S30中，从获取的图像中识别塔筒的边缘。
在一个示例中，可将图像中的指定点作为塔筒的边缘。这里，该指定点可为基于图像捕获器与风力发电机组的塔筒之间的相对关系所确定的塔筒上用于确定塔架净空的点在图像中对应的像素点。也就是说，该指定点可为叶片在运行过程中接触到塔筒可能性最高的位置在图像中对应的像素点。
这里，由于图像捕获器与风力发电机组之间的相对位置关系是固定的，即，风力发电机组的塔筒在所捕获的图像中的位置是固定，那么上述的指定点的位置也是固定的，因此，可以将图像中的指定点作为塔筒的边缘。后续可基于两点之间距离的计算公式得到指定点与叶片的尖端之间的距离来作为塔架净空。以图4所示的图像为例，B点为图像中作为塔筒的边缘的指定点。
在另一示例中，可通过对图像进行边缘检测(或直线检测)的方式来识别塔筒的边缘。
针对上述对用于塔架净空分析的图像进行第二敏感区域提取的情况，可 从第二敏感区域中识别塔筒的边缘。例如，可利用各种直线检测方式从第二敏感区域中检测出直线，将检测出的直线作为塔筒的边缘，但本公开不限于此，也可以采用其他方式来从第二敏感区域中识别出塔筒的边缘，例如，从第二敏感区域中识别预先设置的用于指示塔筒的边缘的标识的方式来确定塔筒的边缘。
下面参照图9来介绍从第二敏感区域中识别塔筒的边缘的步骤。
图9示出根据本公开示例性实施例的识别塔筒的边缘的步骤的流程图。
参照图9，在步骤S301中，从第二敏感区域中提取多个边缘特征点。例如，可将第二敏感区域对应的图像转换为灰度图，从转换后的灰度图中提取边缘特征点。但本公开不限于此，还可以通过其他方式来提取边缘特征点。
在步骤S302中，将提取的多个边缘特征点映射到参数空间中，并基于多个边缘特征点在参数空间绘制相应的多条直线。
例如，考虑点和直线的对应关系，过一点(x 1,y 1)的直线可表示为y 1＝k·x 1+b，将变量和参数互换之后，在已知一个点(x 1,y 1)的情况下，经过这一个点的直线簇可以表示为b＝(-x 1)·k+y 1。也就是说，位于同一条直线上的点具有相同的斜率和截距，反映到参数空间(即，k-b空间)上就是直线簇中的直线会交于同一点(k,b)。
作为示例，假设从第二敏感区域(即，图像空间)中提取出三个边缘特征点(1,1)，(2,2)，(3,3)，这三个边缘特征点在直线y＝1·x+0上，通过变量和参数的互换之后，在参数空间里这三个边缘特征点对应三条直线，即，1＝k+b，2＝2·k+b，3＝3·k+b，这三条直线交于同一点(1,0)，这一点的横纵坐标即为图像空间中的直线的斜率和截距。也就是说，当从参数空间中找到多条直线相交的交点，即可确定在图像空间中的直线。
在一优选示例中，由于上述的变换过程，不能表示直线的斜率为无穷大的情况，因此，可以采用极坐标的方式(例如，Rho＝X·cosθ+Y·sinθ)来表示直线。
在步骤S303中，确定参数空间中的至少一个聚集点。这里，该至少一个聚集点为超过预定数量直线通过的点。
例如，对参数空间中的各直线上的点，我们可采取“投票”(vote)的方法，即累加，当有一条直线经过参数空间内的一点，这一点的得分加1。遍历k-b空间，找出参数空间中累积得分大于或等于预定数值的点作为至少一 个聚集点。
在步骤S304中，根据至少一个聚集点的坐标值确定在第二敏感区域中的至少一条直线，基于至少一条直线确定塔筒的边缘。
例如，根据至少一个聚集点的坐标值确定在第二敏感区域中的至少一条直线的步骤可包括：针对每个聚集点，将该聚集点的横坐标作为直线的斜率，将该聚集点的纵坐标作为直线的截距，得到在第二敏感区域中的与该聚集点对应的直线。也就是说，可以以图像空间的预定坐标系为基准，从第二敏感区域(即，图像空间)中提取边缘特征点，再确定出聚集点之后，在该预定坐标系下绘制相应的直线。
优选地，基于至少一条直线确定塔筒的边缘的步骤可包括：通过对所确定的至少一条直线进行拟合，来基于拟合后得到的拟合直线确定塔筒的边缘。
例如，可从至少一条直线中选取相对距离小于指定距离的直线进行拟合，当拟合后获得一条拟合直线时，将得到的拟合直线作为塔筒的边缘，当拟合后获得两条拟合直线时，将连接两条拟合直线的中点的连线作为塔筒的边缘。
以图4所示的图像为例，基于上述直线检测方法，拟合得到第二敏感区域中直线L1和直线L2时，可将两条直线的中点的连线确定为塔筒的边缘。
以图7所示的图像为例，基于上述直线检测方法，拟合得到第二敏感区域中直线L，将该直线L确定为塔筒的边缘。此时，可利用点到直线的距离计算公式来确定叶片的尖端到塔筒的边缘的距离。
这里，应理解，上述通过对至少一条直线进行拟合来确定塔筒边缘的方式仅为示例，本公开不限于此，其他确定塔筒边缘的方式也是可行的。例如，可计算每条直线的长度，选取长度最长的直线作为塔筒的边缘。
应理解，图9是以从用于塔架净空分析的图像中提取了第二敏感区域为例来介绍从第二敏感区域中识别塔筒的边缘的步骤。但本公开不限于此，图9所示的识别塔筒的边缘的方法也适用于从用于塔架净空分析的图像中识别塔筒的边缘的情况。例如，可从用于塔架净空分析的图像中提取多个边缘特征点以进行塔筒的边缘的识别。
此外，图9所示的识别塔筒的边缘的方式仅为示例，本公开不限于此，本领域技术人员还可以采用其他图像识别方法来从图像中识别出塔筒的边缘。例如，可以从第二敏感区域中提取边缘信息，如提取多个边缘点，通过对多个边缘点进行拟合来得到边缘直线，以作为塔筒的边缘。
返回图1，在步骤S40中，根据确定的叶片的尖端的位置和识别的塔的筒边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。例如，可计算叶片的尖端的位置(如坐标位置)到塔筒的边缘所对应的直线的垂直距离作为塔架净空。
作为示例，根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空的步骤可包括：根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的像素距离；基于预先确定的任意两个像素之间的像素距离与实际距离的对应关系，来利用叶片的尖端到塔筒的边缘的像素距离计算叶片的尖端到塔筒的边缘的实际距离，将所述实际距离确定为塔架净空。这里，任意两个像素可为相邻的两个像素或者图像上指定的两个像素。
图10示出根据本公开示例性实施例的确定风力发电机组的塔架净空的装置的框图。
如图10所示，根据本公开示例性实施例的确定风力发电机组的塔架净空的装置包括图像获取模块10、叶尖检测模块20、塔筒边缘识别模块30和塔架净空确定模块40。
具体说来，图像获取模块10获取风力发电机组在运行过程中的图像。获取的图像中包括风力发电机组的叶片的尖端以及塔筒。
例如，图像获取模块10可获取风力发电机组的叶片在运行过程中的图像，将叶片在运行过程中的图像中包含风力发电机组的叶片的尖端和塔筒的图像确定为用于塔架净空分析的图像。
在一优选实施例中，可利用图像捕获器来捕获风力发电机组的叶片在运行过程中的图像，然后图像捕获器将捕获的图像发送至图像获取模块10。
可以通过合理的设置图像捕获器的安装位置，来使得图像捕获器能够捕获到包含风力发电机组的叶片的尖端和塔筒的图像。下面介绍图像捕获器的两个安装示例。
第一种情况，图像捕获器可被设置在风力发电机组的机舱底部，以捕获到包含风力发电机组的叶片的尖端和塔筒的图像。
第二种情况，图像捕获器可被设置在位于风力发电机组侧面，且与风力发电机组之间的距离为预定距离的指定区域内，以拍摄到包含风力发电机组的叶片的尖端和塔筒的图像。
这里，由于图像捕获器的设置位置是固定的，那么图像捕获器与风力发电机组之间的相对位置关系也是固定的，因此，在该图像捕获器所捕获的图像中哪一区域可能含有塔筒，哪一区域可能含有叶片，也是相对固定。
基于此，优选地，根据本公开示例性实施例的确定风力发电机组的塔架净空的装置可还包括：敏感区域提取模块(图中未示出)，从获取的图像中提取用于检测叶片的尖端的第一敏感区域和用于识别塔筒的边缘的第二敏感区域，后续可以针对提取的第一敏感区域和第二敏感区域进行塔架净空分析。
叶尖检测模块20从获取的图像中确定风力发电机组的叶片的尖端的位置。
针对上述对用于塔架净空分析的图像进行第一敏感区域提取的情况，叶尖检测模块20可以从第一敏感区域中确定风力发电机组的叶片的尖端的位置。
具体说来，叶尖检测模块20可以使用一预定窗口遍历第一敏感区域，当该预定窗口处于第一敏感区域上的任一位置时，叶尖检测模块20可通过以下方式从预定窗口内包含的图像中检测叶片的尖端的位置：
以该任一位置为起点，在第一敏感区域上沿任意方向滑动预定窗口，针对沿每个方向的滑动，确定滑动前与滑动后预定窗口内的像素点的灰度变化程度，并判断灰度变化程度是否满足设定条件，如果针对沿所有方向的滑动，滑动前与滑动后预定窗口内的像素点的灰度变化程度均满足设定条件，则确定预定窗口在该任一位置时包含的图像中存在叶尖特征点，从预定窗口在该任一位置时包含的图像中检测叶尖特征点，并将检测到的叶尖特征点对应的坐标确定为叶片的尖端的位置。
作为示例，叶尖检测模块20可将预定窗口在任一位置时包含的图像中灰度的梯度值最大和/或梯度方向的变化速率最高的像素点确定为叶尖特征点。但本公开不限于此，还可以将任意两条或者两条以上不平行的直线的交点确定为叶尖特征点。
在一优选实施例中，叶尖检测模块20可从第一敏感区域中检测出多个候选叶尖特征点，根据预设条件从多个候选叶尖特征点中确定最终的叶尖特征点。
叶尖检测模块20根据用于捕获用于塔架净空分析的图像的图像捕获器与风力发电机组之间的相对位置关系，从多个候选叶尖特征点中选择最靠近 地面的候选叶尖特征点作为最终的叶尖特征点。
针对图像捕获器被设置在机舱底部的情况，叶尖检测模块20可将多个候选叶尖特征点中的Y轴坐标值最大的点确定为最终的叶尖特征点。
针对图像捕获器被设置在指定区域内时，叶尖检测模块20可将多个候选叶尖特征点中的Y轴坐标值最小的点确定为最终的叶尖特征点。
塔筒边缘识别模块30从获取的图像中识别塔筒的边缘。
例如，针对上述对用于塔架净空分析的图像进行第二敏感区域提取的情况，塔筒边缘识别模块30可从第二敏感区域中识别塔筒的边缘。
在一个示例中，塔筒边缘识别模块30可将图像中的指定点作为塔筒的边缘。这里，该指定点可为基于图像捕获器与风力发电机组的塔筒之间的相对关系所确定的塔筒上用于确定塔架净空的点在图像中对应的像素点。也就是说，该指定点可为叶片在运行过程中接触到塔筒可能性最高的位置在图像中对应的像素点。
在另一示例中，塔筒边缘识别模块30可通过对图像进行边缘检测的方式来识别塔筒的边缘。
塔筒边缘识别模块30可利用各种图像识别方法来从第二敏感区域中识别塔筒的边缘。下面参照图11来介绍塔筒边缘识别模块30的功能。应理解，图11所示的识别塔筒的边缘的方式仅为一优选示例，其他用于识别塔筒的边缘的图像识别方式也是可行的。
图11示出根据本公开示例性实施例的塔筒边缘识别模块30的框图。
如图11所示，根据本公开示例性实施例的塔筒边缘识别模块30可包括特征提取子模块301、转换子模块302、聚集点确定子模块303和边缘确定子模块304。
具体说来，特征提取子模块301从第二敏感区域中提取多个边缘特征点。例如，特征提取子模块301可将第二敏感区域对应的图像转换为灰度图，从转换后的灰度图中提取边缘特征点。
转换子模块302将提取的多个边缘特征点映射到参数空间中，并基于多个边缘特征点在参数空间绘制相应的多条直线。
由于已经在图9的步骤302中描述了转换子模块302获得多条直线的详细过程，本公开对此部分的内容不再赘述。
聚集点确定子模块303确定参数空间中的至少一个聚集点。这里，该至 少一个聚集点为超过预定数量直线通过的点。
由于已经在图9的步骤303中描述了聚集点确定子模块303确定至少一个聚集点的详细过程，本公开对此部分的内容不再赘述。
边缘确定子模块304根据至少一个聚集点的坐标值确定在第二敏感区域中的至少一条直线，基于至少一条直线确定塔筒的边缘。
例如，针对每个聚集点，边缘确定子模块304可将该聚集点的横坐标作为直线的斜率，将该聚集点的纵坐标作为直线的截距，得到在第二敏感区域中与该聚集点对应的直线。
例如，边缘确定子模块304可从至少一条直线中选取相对距离小于指定距离的直线进行拟合，当拟合后获得一条拟合直线时，将得到的拟合直线作为塔筒的边缘，当拟合后获得两条拟合直线时，将连接两条拟合直线的中点的连线作为塔筒的边缘。
返回图10，塔架净空确定模块40根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。例如，塔架净空确定模块40可计算叶片的尖端的位置到塔筒的边缘所对应的直线的垂直距离作为塔架净空。
塔架净空确定模块40可根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的像素距离，基于预先确定的任意两个像素之间的像素距离与实际距离的对应关系，来利用叶片的尖端到塔筒的边缘的像素距离计算叶片的尖端到塔筒的边缘的实际距离，将所述实际距离确定为塔架净空。
根据本公开的示例性实施例还提供一种存储有计算机程序的计算机可读存储介质。该计算机可读存储介质存储有当被处理器执行时使得处理器执行上述确定风力发电机组的塔架净空的方法的计算机程序。该计算机可读记录介质是可存储由计算机系统读出的数据的任意数据存储装置。计算机可读记录介质的示例包括：只读存储器、随机存取存储器、只读光盘、磁带、软盘、光数据存储装置和载波(诸如经有线或无线传输路径通过互联网的数据传输)。
采用本公开示例性实施例的确定风力发电机组的塔架净空的方法和装置，能够实现实时监测风力发电机组的塔架净空，以有效避免叶片扫塔带来的损失。
此外，采用本公开示例性实施例的确定风力发电机组的塔架净空的方法 和装置，通过合理设计用于支撑图像捕获器的支架以及合理选取图像捕获器的安装位置，不仅能完全实现塔架净空视频监控的功能并且还能保证风力发电机组的安全运行。
此外，采用本公开示例性实施例的确定风力发电机组的塔架净空的方法和装置，可以相对简单地获取到风力发电机组的塔架净空，无需人工测量，方便快捷。
此外，采用本公开示例性实施例的确定风力发电机组的塔架净空的方法和装置，基于叶尖特征点检测方法来检测叶片的尖端，还可通过直线检测方式识别塔筒的边缘，进而得到叶片的尖端与塔筒的边缘之间的距离。
此外，采用本公开示例性实施例的确定风力发电机组的塔架净空的方法和装置，使用单目视觉技术实现对于塔架净空高精度的测量。
尽管已参照优选实施例表示和描述了本公开，但本领域技术人员应该理解，在不脱离由权利要求限定的本公开的精神和范围的情况下，可以对这些实施例进行各种修改和变换。

Claims (13)

  1. 一种确定风力发电机组的塔架净空的方法，所述方法包括：
    获取风力发电机组在运行过程中的图像，所述图像包括所述风力发电机组的叶片的尖端以及塔筒；
    从获取的图像中确定所述风力发电机组的叶片的尖端的位置；
    从获取的图像中识别所述塔筒的边缘；
    根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
  2. 如权利要求1所述的方法，其中，所述方法还包括：从获取的图像中提取用于检测叶片的尖端的第一敏感区域和用于识别塔筒的边缘的第二敏感区域，
    其中，从第一敏感区域中确定所述风力发电机组的叶片的尖端的位置，从第二敏感区域中识别所述塔筒的边缘。
  3. 如权利要求2所述的方法，其中，从第一敏感区域中确定所述风力发电机组的叶片的尖端的位置包括：
    使用一预定窗口遍历第一敏感区域，当所述预定窗口处于第一敏感区域上的任一位置时，通过以下方式从所述预定窗口内包含的图像中检测叶片的尖端的位置：
    以所述任一位置为起点，在第一敏感区域上沿任意方向滑动所述预定窗口；
    针对沿每个方向的滑动，确定滑动前与滑动后所述预定窗口内的像素点的灰度变化程度，并判断所述灰度变化程度是否满足设定条件；
    如果针对沿所有方向的滑动，滑动前与滑动后所述预定窗口内的像素点的灰度变化程度均满足设定条件，则确定所述预定窗口在所述任一位置时包含的图像中存在叶尖特征点；
    从所述预定窗口在所述任一位置时包含的图像中检测叶尖特征点，并将检测到的叶尖特征点对应的坐标确定为叶片的尖端的位置。
  4. 如权利要求3所述的方法，其中，从所述预定窗口在所述任一位置时包含的图像中检测叶尖特征点包括：
    将所述预定窗口在所述任一位置时包含的图像中灰度的梯度值和/或梯 度方向的变化速率最高的像素点确定为叶尖特征点。
  5. 如权利要求2所述的方法，其中，从第一敏感区域中确定所述风力发电机组的叶片的尖端的位置包括：
    从第一敏感区域中检测出多个候选叶尖特征点；
    根据预设条件从所述多个候选叶尖特征点中确定最终的叶尖特征点。
  6. 如权利要求5所述的方法，其中，用于捕获风力发电机组在运行过程中的图像的图像捕获器被设置在风力发电机组的机舱底部，或者所述图像捕获器被设置在位于风力发电机组侧面且与风力发电机组之间的距离为预定距离的指定区域内，
    其中，根据预设条件从所述多个候选叶尖特征点中确定最终的叶尖特征点包括：
    当所述图像捕获器被设置在机舱底部时，将所述多个候选叶尖特征点中的Y轴坐标值最大的点确定为最终的叶尖特征点；
    当所述图像捕获器被设置在所述指定区域内时，将所述多个候选叶尖特征点中的Y轴坐标值最小的点确定为最终的叶尖特征点。
  7. 如权利要求2所述的方法，其中，从第二敏感区域中识别所述塔筒的边缘包括：
    从第二敏感区域中提取多个边缘特征点；
    将所述多个边缘特征点映射到参数空间中，并基于所述多个边缘特征点在参数空间绘制对应的多条直线；
    确定参数空间中的至少一个聚集点，所述至少一个聚集点为超过预定数量直线通过的点；
    根据所述至少一个聚集点的坐标值确定在第二敏感区域中的至少一条直线，基于所述至少一条直线确定塔筒的边缘。
  8. 如权利要求7所述的方法，其中，根据所述至少一个聚集点的坐标值确定在第二敏感区域中的至少一条直线包括：
    针对每个聚集点，将该聚集点的横坐标作为直线的斜率，将该聚集点的纵坐标作为直线的截距，得到在第二敏感区域中的与该聚集点对应的直线。
  9. 如权利要求7所述的方法，其中，基于所述至少一条直线确定塔筒的边缘包括：
    当对所述至少一条直线进行拟合获得一条拟合直线时，将得到的拟合直 线作为塔筒的边缘；或者，
    当对所述至少一条直线进行拟合获得两条拟合直线时，将连接所述两条拟合直线的中点的连线作为塔筒的边缘。
  10. 如权利要求1所述的方法，其中，根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空包括：
    根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的像素距离；
    基于预先确定的任意两个像素之间的像素距离与实际距离的对应关系，来利用叶片的尖端到塔筒的边缘的像素距离计算叶片的尖端到塔筒的边缘的实际距离，将所述实际距离确定为塔架净空。
  11. 一种确定风力发电机组的塔架净空的装置，所述装置包括：
    图像获取模块，被配置为获取风力发电机组在运行过程中的图像，所述图像包括所述风力发电机组的叶片的尖端以及塔筒；
    叶尖检测模块，被配置为从获取的图像中确定所述风力发电机组的叶片的尖端的位置；
    塔筒边缘识别模块，被配置为从获取的图像中识别所述塔筒的边缘；
    塔架净空确定模块，被配置为根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
  12. 一种塔架净空监测系统，所述塔架净空监测系统包括：
    图像捕获器，用于捕获风力发电机组的叶片在运行过程中的图像；
    处理器，被配置为：
    从所捕获的图像中获取包括所述风力发电机组的叶片的尖端以及塔筒的图像；
    从获取的图像中确定所述风力发电机组的叶片的尖端的位置；
    从获取的图像中识别所述塔筒的边缘；
    根据确定的叶片的尖端的位置和识别的塔筒的边缘，计算叶片的尖端到塔筒的边缘的距离以获得塔架净空。
  13. 一种存储有计算机程序的计算机可读存储介质，当所述计算机程序在被处理器执行时实现如权利要求1至10中任意一项所述的确定风力发电机组的塔架净空的方法。
PCT/CN2019/109391 2018-11-30 2019-09-30 确定风力发电机组的塔架净空的方法和装置 WO2020108088A1 (zh)

Priority Applications (3)

Application Number Priority Date Filing Date Title
EP19890997.0A EP3744974A4 (en) 2018-11-30 2019-09-30 METHOD AND DEVICE FOR DETERMINING THE TOWER DISTANCE FOR A WIND TURBINE
US16/976,403 US11421659B2 (en) 2018-11-30 2019-09-30 Method and device for determining tower clearance for wind turbine
AU2019390462A AU2019390462B2 (en) 2018-11-30 2019-09-30 Method and device for determining tower clearance for wind turbine

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811454363.7 2018-11-30
CN201811454363.7A CN111255636B (zh) 2018-11-30 2018-11-30 确定风力发电机组的塔架净空的方法和装置

Publications (1)

Publication Number Publication Date
WO2020108088A1 true WO2020108088A1 (zh) 2020-06-04

Family

ID=70853791

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/109391 WO2020108088A1 (zh) 2018-11-30 2019-09-30 确定风力发电机组的塔架净空的方法和装置

Country Status (5)

Country Link
US (1) US11421659B2 (zh)
EP (1) EP3744974A4 (zh)
CN (1) CN111255636B (zh)
AU (1) AU2019390462B2 (zh)
WO (1) WO2020108088A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112558632A (zh) * 2020-12-03 2021-03-26 上海扩博智能技术有限公司 无人机巡检路径转换方法、系统、设备和存储介质

Families Citing this family (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113864132B (zh) * 2020-06-30 2023-04-07 乌鲁木齐金风天翼风电有限公司 塔架净空的监测方法、装置及设备
CN113915074A (zh) * 2020-07-10 2022-01-11 深圳市镭神智能系统有限公司 风力发电机的监控方法、系统及风力发电系统
CN111878319B (zh) * 2020-07-13 2021-07-16 明阳智慧能源集团股份公司 一种基于多激光头的风机叶片净空自动监测方法及系统
CN111654642B (zh) * 2020-07-22 2021-10-01 上海扩博智能技术有限公司 拍摄风机叶片的曝光值调整方法、系统、设备和存储介质
CN111911364B (zh) * 2020-09-11 2021-11-23 上海电气风电集团股份有限公司 叶尖塔筒净空监测方法
CN111980872B (zh) * 2020-09-18 2022-06-14 南京牧镭激光科技有限公司 一种测量风力发电机叶片到塔筒的距离传感装置
CN112883503B (zh) * 2020-11-05 2021-08-03 中国长江三峡集团有限公司 基于ptfe膜对风机叶片气动特性影响的数值模拟方法
CN112761897B (zh) * 2021-01-13 2022-06-14 国电联合动力技术有限公司 风电机组叶片监控方法、装置及风力发电机
CN113309674B (zh) * 2021-03-31 2022-07-15 新疆金风科技股份有限公司 风力发电机组的净空距离确定方法及装置
CN113309673A (zh) * 2021-06-29 2021-08-27 新疆金风科技股份有限公司 风力发电机组的静态净空距离确定方法和装置
CN115726934A (zh) * 2021-08-31 2023-03-03 北京金风科创风电设备有限公司 一种风力发电机净空值的测量方法及装置
WO2023188378A1 (ja) * 2022-03-31 2023-10-05 株式会社Acsl 無人航空機を用いて風力発電装置のロータ方向・アジマス角を推定するためのシステム、方法、プログラム及びプログラムを記憶した記憶媒体
CN117272550B (zh) * 2023-11-17 2024-03-22 陕西空天信息技术有限公司 一种流道设计的方法、装置、设备及计算机存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120200699A1 (en) * 2011-02-08 2012-08-09 Steffen Bunge Balancing of Wind Turbine Parts
CN103982378A (zh) * 2014-04-25 2014-08-13 广东电网公司电力科学研究院 基于机器视觉图像对电力系统风力发电机叶片表面结冰故障进行诊断的方法
CN106286152A (zh) * 2016-09-14 2017-01-04 北京金风科创风电设备有限公司 风力发电机组的叶片状态监测装置及监测方法
CN106762451A (zh) * 2016-12-05 2017-05-31 北京金风科创风电设备有限公司 基于无人机的风机叶片损伤检测方法、装置及系统
CN207598432U (zh) * 2017-11-30 2018-07-10 远景能源（江苏）有限公司 监测风力发电机叶片弯度的装置
CN108506172A (zh) * 2018-06-01 2018-09-07 河北工业大学 一种风力发电叶片健康状态在线检测系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110135466A1 (en) * 2010-01-14 2011-06-09 General Electric Company System and method for monitoring and controlling wind turbine blade deflection
CN102434403B (zh) * 2010-09-29 2015-09-09 通用电气公司 用于风力涡轮机检查的系统及方法
GB201222540D0 (en) * 2012-12-14 2013-01-30 Lm Wp Patent Holding As A system and method for wind turbine sensor calibration
US9453500B2 (en) * 2013-03-15 2016-09-27 Digital Wind Systems, Inc. Method and apparatus for remote feature measurement in distorted images
US10378517B2 (en) * 2014-03-04 2019-08-13 Steffen Bunge Method for replacing the blades of a wind turbine to maintain safe operation
CN108799011B (zh) * 2017-04-28 2020-01-31 北京金风科创风电设备有限公司 对风电机组的叶片进行监测的设备和方法
CN111433455B (zh) * 2017-12-04 2022-08-09 日本电产Ssb风系统有限公司 风力涡轮机塔架至桨叶末端测量系统
CN111255637A (zh) * 2018-11-30 2020-06-09 北京金风科创风电设备有限公司 塔架净空实时监测系统及其方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120200699A1 (en) * 2011-02-08 2012-08-09 Steffen Bunge Balancing of Wind Turbine Parts
CN103982378A (zh) * 2014-04-25 2014-08-13 广东电网公司电力科学研究院 基于机器视觉图像对电力系统风力发电机叶片表面结冰故障进行诊断的方法
CN106286152A (zh) * 2016-09-14 2017-01-04 北京金风科创风电设备有限公司 风力发电机组的叶片状态监测装置及监测方法
CN106762451A (zh) * 2016-12-05 2017-05-31 北京金风科创风电设备有限公司 基于无人机的风机叶片损伤检测方法、装置及系统
CN207598432U (zh) * 2017-11-30 2018-07-10 远景能源（江苏）有限公司 监测风力发电机叶片弯度的装置
CN108506172A (zh) * 2018-06-01 2018-09-07 河北工业大学 一种风力发电叶片健康状态在线检测系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3744974A4 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112558632A (zh) * 2020-12-03 2021-03-26 上海扩博智能技术有限公司 无人机巡检路径转换方法、系统、设备和存储介质

Also Published As

Publication number Publication date
EP3744974A4 (en) 2021-08-18
US11421659B2 (en) 2022-08-23
EP3744974A1 (en) 2020-12-02
CN111255636B (zh) 2023-07-25
US20210003114A1 (en) 2021-01-07
AU2019390462A1 (en) 2020-08-27
CN111255636A (zh) 2020-06-09
AU2019390462B2 (en) 2022-01-20

Similar Documents

Publication Publication Date Title
WO2020108088A1 (zh) 确定风力发电机组的塔架净空的方法和装置
CN106525025B (zh) 一种变电站巡检机器人路线规划导航方法
CN109443480B (zh) 基于图像处理的水位标尺定位及水位测量方法
CN111855664B (zh) 一种可调节隧道病害三维检测系统
CN107462223B (zh) 一种公路转弯前行车视距自动测量装置及测量方法
CN108416791A (zh) 一种基于双目视觉的并联机构动平台位姿监测与跟踪方法
CN109559354A (zh) 测量塔架净空的方法、双目系统的标定方法以及装置
CN105373135A (zh) 一种基于机器视觉的飞机入坞引导和机型识别的方法及系统
WO2019119659A1 (zh) 用于风力发电机组的涡激振动的监测方法和设备
CN104700395A (zh) 一种构造物外观裂缝检测方法及系统
CN110533649B (zh) 一种无人机通用结构裂缝识别检测装置及方法
US20230105991A1 (en) Method of imaging a wind turbine rotor blade
CN111243003A (zh) 车载双目摄像机及其检测道路限高杆的方法、装置
CN109492525B (zh) 一种测量基站天线工程参数的方法
Dillabaugh et al. Semi-automated extraction of rivers from digital imagery
CN106403901A (zh) 一种浮标姿态测量装置及方法
CN111832760B (zh) 一种基于视觉算法的井盖自动巡检方法
Guo et al. Detecting and positioning of wind turbine blade tips for uav-based automatic inspection
CN103913149A (zh) 一种基于stm32单片机的双目测距系统及其测距方法
CN106600561A (zh) 基于投影映射的航拍图像透视畸变自动校正方法
CN115511878A (zh) 一种边坡地表位移监测方法、装置、介质及设备
CN115841633A (zh) 一种电力塔和电力线关联矫正的电力塔和电力线检测方法
CN102054174A (zh) 用于定位车牌的方法和系统
CN107767366B (zh) 一种输电线路拟合方法及装置
CN116486212A (zh) 一种基于计算机视觉的水尺识别方法、系统及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19890997

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2019390462

Country of ref document: AU

Date of ref document: 20190930

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2019890997

Country of ref document: EP

Effective date: 20200827

WWE Wipo information: entry into national phase

Ref document number: 2019890997

Country of ref document: EP

NENP Non-entry into the national phase

Ref country code: DE