Device and method for monitoring and evaluating collision risk of offshore wind turbine blade to birds

Abstract

The invention discloses a device and method for monitoring and evaluating the collision risk of offshore wind turbine blades to birds, and the device comprises a reading module which is used for reading a monitoring image; the recognition module is used for performing feature extraction on a target in the monitoring image through a preset convolutional neural network to obtain a recognition result; the analysis module is used for framing a fan blade operation area based on the monitoring image and the recognition result so as to obtain a fan blade movement track; marking the specific position of the bird in each frame of image through a preset target detection model, and performing line connection to form a bird flight path; the evaluation module is used for judging the behavior type of the birds dealing with blade rotation based on the intersection condition of the bird flight track and the fan blade motion track; and calculating the corresponding proportion of the monitored birds in the behavior type, and obtaining corresponding evaluation information according to the obtained proportion data. The method can overcome the defects of time and labor consumption, large error and imperfect evaluation system in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore wind farms, and in particular to a device and method for monitoring and evaluating the risk of bird collisions with offshore wind turbine blades. Background Art

[0002] At present, collision is the most concerned impact of offshore wind turbine operation on birds. It may interfere with bird migration, foraging and other behavioral activities to a certain extent. Especially during the peak period of bird migration, the high-speed rotation of the blades may cause bird injuries or even death, thus affecting the ecological balance. Therefore, scientifically and accurately assessing the actual impact of wind turbine blade rotation on nearby birds is of great significance for operators to optimize wind turbine blade design and take effective intervention measures.

[0003] In the current existing technologies, most of the solutions adopted are based on numerical models or indirectly assessing the risk of collision between birds and wind turbine blades by counting the number of dead birds in wind farms. However, these two methods are not only time-consuming and labor-intensive, but also have large errors. They are also unable to effectively track the interaction between birds and wind turbine blades, making it difficult to accurately assess the actual impact of wind turbine blade rotation on birds. Summary of the invention

[0004] In view of the technical defects mentioned in the background technology, the purpose of the embodiments of the present invention is to provide a monitoring and evaluation device and method for the risk of bird collision of offshore wind turbine blades, so as to solve the defects of the existing monitoring and evaluation methods proposed in the above background technology, which are time-consuming, labor-intensive and have large errors.

[0005] To achieve the above objectives, in a first aspect, an embodiment of the present invention provides a monitoring and evaluation device for the risk of bird collision with offshore wind turbine blades, the monitoring and evaluation device comprising:

[0006] A reading module, used for reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on the wind turbine tower;

[0007] A recognition module, used to extract features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result and perform labeling; wherein the recognition result includes a fan blade and a target bird;

[0008] Analysis modules for:

[0009] Based on the monitoring image and the recognition result, the operation area of ​​the fan blade is framed to obtain the movement trajectory of the fan blade;

[0010] The specific location of the bird is marked in each frame image through the preset target detection model, and lines are connected to form the corresponding bird flight trajectory;

[0011] Evaluation modules for:

[0012] Based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory, determining the bird's behavior type in response to blade rotation;

[0013] The corresponding proportions of the monitored birds in the behavior types are calculated, and corresponding evaluation information is obtained based on the obtained proportion data.

[0014] As a specific implementation of the present application, the identification module is also used for:

[0015] The recognition result is optimized using a preset multi-layer neural network to clarify the species corresponding to the target bird.

[0016] As a specific implementation of the present application, obtaining the motion trajectory of the fan blades specifically includes:

[0017] Using the cv2.rectangle function provided by OpenCV to draw a rectangular frame on the monitoring image, mark the running track of the wind blade on the monitoring image, and obtain the coordinates of the rectangular frame;

[0018] The width, height and area of ​​the rectangular frame are determined based on the obtained coordinates and stored for subsequent determination of whether the flight trajectory of the bird overlaps with the operation area of ​​the wind blade.

[0019] As an optimized implementation of the present application, the analysis module is also used for:

[0020] The Kalman filter is used to estimate and predict the motion state of the target bird to improve the stability and accuracy of tracking.

[0021] As an optimized implementation of the present application, the analysis module is also used for:

[0022] Through the data association algorithm, the position, size and grayscale features of the target birds in multiple frames are matched so that the detected target birds can be associated with the subsequent flight trajectory to ensure that the trajectory of each bird is independent and continuous.

[0023] As a specific implementation of the present application, the behavior types include adjusted behavior, unadjusted behavior and collision phenomenon;

[0024] When a bird suddenly adjusts its direction and returns or stops flying when approaching a blade, it is considered that the bird exhibits the said adjustment behavior when the blade rotates;

[0025] When the bird flies in a direction parallel to the blade, or turns multiple times within the blade's sweep area to avoid the blade, the bird is considered to exhibit the described adjustment behavior in response to the blade's rotation;

[0026] When the bird enters the blade sweep area but the flight trajectory does not change, the bird is considered to exhibit the unadjusted behavior;

[0027] When a bird collides with a fan blade, it is considered that the bird and the blade have the collision phenomenon.

[0028] In a second aspect, an embodiment of the present invention further provides a method for monitoring and evaluating the risk of bird collision of offshore wind turbine blades, which is applied to the device for monitoring and evaluating the risk of bird collision of offshore wind turbine blades described in the first aspect, and the method comprises the following steps:

[0029] Reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on the wind turbine tower;

[0030] Extracting features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result and label it; wherein the recognition result includes a fan blade and a target bird;

[0031] Based on the monitoring image and the recognition result, the operation area of ​​the fan blade is framed to obtain the movement trajectory of the fan blade;

[0032] The specific location of the bird is marked in each frame image through the preset target detection model, and lines are connected to form the corresponding bird flight trajectory;

[0033] Based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory, determining the bird's behavior type in response to blade rotation;

[0034] The corresponding proportions of the monitored birds in the behavior types are calculated, and corresponding evaluation information is obtained based on the obtained proportion data.

[0035] The technical solution provided by the embodiment of the present invention can monitor the behavior dynamics of birds around offshore wind turbines, effectively identify and track the flight paths of birds; through monitoring, the behavior changes of birds when passing through wind turbine blades can be discovered and mastered in a timely manner, providing effective data support for subsequent scientific assessment of bird threat risks; the defects of the prior art, such as time-consuming and labor-intensive, large errors, and inability to effectively track the interactive behaviors of birds and wind turbine blades, are overcome;

[0036] Through accurate data recording and analysis, it can help ecological and environmental departments and wind farm managers formulate scientific protection measures, reduce the interference and harm of wind turbines to birds, promote the coordinated development of offshore wind power and marine environmental protection, and provide a basis for making scientific decisions to reduce the negative impact of the ecological environment. At the same time, it also provides more ideas for future offshore wind power planning and project environmental assessments;

[0037] It can also improve the operational efficiency of wind farms and avoid equipment damage and downtime for maintenance caused by bird collisions, thereby improving the operating efficiency and reliability of wind turbines. At the same time, it can assist wind farm managers in optimizing the layout and operation strategies of wind turbines, minimize the impact on bird migration, and achieve the goal of sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the drawings required for use in the specific implementation or the description of the prior art are briefly introduced below.

[0039] Figure 1 This is a principle block diagram of a device for monitoring and evaluating the risk of bird collision with offshore wind turbine blades provided by an embodiment of the present invention;

[0040] Figure 2 is a schematic diagram of an adjustment behavior provided by an embodiment of the present invention;

[0041] Figure 3 is a schematic diagram of another adjustment behavior provided by an embodiment of the present invention;

[0042] Figure 4 is a schematic diagram of an unadjusted behavior provided by an embodiment of the present invention;

[0043] Figure 5 is a schematic diagram of a collision phenomenon provided by an embodiment of the present invention;

[0044] Figure 6 It is a flow chart of a method for monitoring and evaluating the risk of bird collision with offshore wind turbine blades provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0047] As used in this specification and the appended claims, the term "if" may be interpreted as "when" or "upon" or "in response to determining" or "in response to detecting," depending on the context. Similarly, the phrases "if it is determined" or "if [described condition or event] is detected" may be interpreted as meaning "upon determination" or "in response to determining" or "upon detection of [described condition or event]" or "in response to detecting [described condition or event]," depending on the context.

[0048] It should be noted that, unless otherwise specified, the technical terms or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0049] Please refer to Figure 1 , an embodiment of the present invention provides a monitoring and evaluation device for the risk of bird collision of offshore wind turbine blades, the monitoring and evaluation device comprising:

[0050] A reading module, used for reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on the wind turbine tower;

[0051] A recognition module, used to extract features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result and perform labeling; wherein the recognition result includes a fan blade and a target bird;

[0052] Analysis modules for:

[0053] Based on the monitoring image and the recognition result, the operation area of ​​the fan blade is framed to obtain the movement trajectory of the fan blade;

[0054] The specific location of the bird is marked in each frame through the preset target detection model, and connected to form the corresponding bird flight trajectory;

[0055] Evaluation modules for:

[0056] Based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory, determining the bird's behavior type in response to blade rotation;

[0057] The corresponding proportions of the monitored birds in the behavior types are calculated, and corresponding evaluation information is obtained based on the obtained proportion data.

[0058] In this embodiment, the monitoring device is built on the wind turbine tower, with an infrared triggered camera as the front-end acquisition device. The device is charged by external solar energy, and the infrared sensor is turned on at a fixed time to capture images of birds. Wind turbines in the edge area of ​​the wind farm are preferentially selected as monitoring points, and fixed brackets, solar panels, infrared triggered cameras, waterproof housings and locking parts are installed in sequence at a height of 1-2m on the wind turbine tower to complete the bird monitoring device, and shoot from bottom to top.

[0059] At the same time, adjust the camera shooting angle, magnification, zoom and other parameters to ensure that the picture covers the entire movement trajectory of the wind blades and has enough space to capture the flight trajectory of birds.

[0060] Each frame of the monitoring image captured is also preprocessed to enhance image quality and remove noise.

[0061] In order to achieve the contextual association of each frame of the monitoring image to ensure the accuracy and real-time performance of the recognition, the recognition module is also used to:

[0062] The recognition result is optimized using a preset multi-layer neural network to clarify the species corresponding to the target bird.

[0063] It should be noted that a multi-layer neural network is composed of multiple layer structures, and each layer is composed of a number of neuron nodes. Any node in this layer is connected to every node in the previous layer, and they provide input. The output of the node is generated after calculation and serves as the input of the nodes in the next layer.

[0064] In this embodiment, the analysis module deeply analyzes the screen image of the captured video to obtain the operation area of ​​the wind blade;

[0065] The obtaining of the motion trajectory of the fan blades specifically includes:

[0066] Using the cv2.rectangle function provided by OpenCV to draw a rectangular frame on the monitoring image, mark the running track of the wind blade on the monitoring image, and obtain the coordinates of the rectangular frame;

[0067] The width, height and area of ​​the rectangular frame are determined based on the obtained coordinates and stored for subsequent determination of whether the flight trajectory of the bird overlaps with the operation area of ​​the wind blade.

[0068] Specifically, through this function, the running track of the wind blade can be manually or automatically marked on the image, and the coordinates of the rectangular frame can be obtained. In OpenCV, the coordinates of the rectangular frame are usually determined by two points: the coordinates of the starting point of the upper left corner (x1, y1) and the coordinates of the end point of the lower right corner (x2, y2). Through these two points, the boundaries of the rectangular frame can be determined, that is, the width of the rectangle = |x2-x1|, the height of the rectangle = |y2-y1|, and the area of ​​the rectangular frame = |x2-x1|×|y2-y1|. The obtained rectangular frame coordinates can be directly stored in the database as the basic data for comparison with the bird's flight trajectory in the subsequent steps. The area of ​​the rectangular frame can be used to determine whether the bird's flight trajectory overlaps with the wind blade operation area.

[0069] It should be noted that the preset target detection model is built with a convolutional neural network, and uses a convolutional neural network in combination with the YOLO algorithm to extract image features, as well as object classification and bounding box regression through a fully connected layer; it detects the position of the bird in each frame, outputs the bounding box or center point coordinates of the bird, and marks the specific position of the bird with a dot in the picture; these coordinates will be used as input data for trajectory tracking to achieve subsequent real-time tracking of the bird; and local features such as bird edges and textures are extracted from each image through multi-layer convolution operations.

[0070] Furthermore, the analysis module is also used for:

[0071] The Kalman filter is used to estimate and predict the motion state of the target bird to improve the stability and accuracy of tracking. This can smooth and predict the flight trajectory of the bird, estimate and predict the motion state (such as position and speed) of the target bird in the case of high noise, and improve the stability and accuracy of tracking.

[0072] In this embodiment, the analysis module is also used for:

[0073] Through the data association algorithm, the position, size and grayscale features of the target birds in multiple frames are matched so that the detected target birds can be associated with the subsequent flight trajectory to ensure that the trajectory of each bird is independent and continuous.

[0074] When applied, the data association algorithm can adopt probabilistic data association, Hungarian algorithm, etc.; the size and grayscale features are local features extracted by the aforementioned multi-layer convolution operation; through the above processing, the multi-target tracking problem (i.e., the situation where multiple birds appear at the same time) is solved.

[0075] Bird flight trajectory drawing. In each frame of the image, a small dot is drawn at the current position of the bird (such as using the cv2.circle function in OpenCV) and these dots are connected to form a bird flight trajectory diagram. As time goes by, connecting these dots can clearly show the flight path of the bird. After the real-time tracking is completed, the coordinate information of the trajectory point is stored in the database and the coincidence with the above-mentioned wind blade running trajectory is judged to facilitate further evaluation of the impact of wind turbine blades on bird flight.

[0076] In this embodiment, refer to Figures 2 to 5 , the behavior types include adjusted behavior, unadjusted behavior and collision phenomenon; wherein, Figure 2 To prevent birds from suddenly changing direction when approaching the leaves; Figure 3 For birds flying in a direction parallel to the blades; Figure 4 The bird enters the blade sweep area and the flight trajectory does not change; Figure 5 Birds collide with wind turbine blades;

[0077] When a bird suddenly adjusts its direction and returns or stops flying when approaching a blade, it is considered that the bird exhibits the said adjustment behavior when the blade rotates;

[0078] When the bird flies in a direction parallel to the blade, or turns multiple times within the blade's sweep area to avoid the blade, the bird is considered to exhibit the described adjustment behavior in response to the blade's rotation;

[0079] When the bird enters the blade sweep area but the flight trajectory does not change, the bird is considered to exhibit the unadjusted behavior;

[0080] When a bird collides with a fan blade, it is considered that the bird and the blade have the collision phenomenon.

[0081] The proportion of birds that showed "adjustment" behavior, did not adjust" behavior and collided among the monitored birds was calculated. When the proportion of birds that showed "adjustment" behavior was higher than 90% and the proportion of birds that collided was lower than 2%, the risk of collision between birds and leaves was considered low; when the proportion of birds that showed "adjustment" behavior was between 70% and 90% and the proportion of birds that collided was lower than 5%, the risk of collision between birds and leaves was considered moderate; when the proportion of birds that collided was higher than 5% or the proportion of birds that showed "adjustment" behavior was lower than 70%, the risk of collision between birds and leaves was considered high.

[0082] The above scheme can monitor the dynamic behavior of birds around offshore wind turbines, effectively identify and track the flight paths of birds; through monitoring, the behavioral changes of birds when passing through wind turbine blades can be discovered and grasped in time, providing effective data support for subsequent scientific assessment of bird threat risks; it overcomes the defects of existing technologies that are time-consuming and labor-intensive, have large errors, and cannot effectively track the interactive behaviors of birds and wind turbine blades;

[0083] Through accurate data recording and analysis, it can help ecological and environmental departments and wind farm managers formulate scientific protection measures, reduce the interference and harm of wind turbines to birds, promote the coordinated development of offshore wind power and marine environmental protection, and provide a basis for making scientific decisions to reduce the negative impact of the ecological environment. At the same time, it also provides more ideas for future offshore wind power planning and project environmental assessments;

[0084] It can also improve the operational efficiency of wind farms and avoid equipment damage and downtime for maintenance caused by bird collisions, thereby improving the operating efficiency and reliability of wind turbines. At the same time, it can assist wind farm managers in optimizing the layout and operation strategies of wind turbines, minimize the impact on bird migration, and achieve the goal of sustainable development.

[0085] Reference Figure 6 Based on the same inventive concept, an embodiment of the present invention further provides a method for monitoring and evaluating the risk of bird collision of offshore wind turbine blades, which is applied to a device for monitoring and evaluating the risk of bird collision of offshore wind turbine blades as described in the first aspect, and the method comprises the following steps:

[0086] S101, reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on a wind turbine tower;

[0087] S102, extracting features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result, and labeling the result; wherein the recognition result includes a fan blade and a target bird;

[0088] S103, framing the operation area of ​​the fan blade based on the monitoring image and the recognition result to obtain the movement trajectory of the fan blade;

[0089] S104, marking the specific position of the bird in each frame image by using a preset target detection model, and connecting lines to form a corresponding bird flight trajectory;

[0090] S105, judging the behavior type of the bird in response to blade rotation based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory;

[0091] S106, calculating the corresponding proportion of the monitored birds in the behavior type, and obtaining corresponding evaluation information according to the obtained proportion data.

[0092] Furthermore, the method further comprises:

[0093] The recognition result is optimized using a preset multi-layer neural network to clarify the species corresponding to the target bird.

[0094] The obtaining of the motion trajectory of the fan blades specifically includes:

[0095] Using the cv2.rectangle function provided by OpenCV to draw a rectangular frame on the monitoring image, mark the running track of the wind blade on the monitoring image, and obtain the coordinates of the rectangular frame;

[0096] The width, height and area of ​​the rectangular frame are determined based on the obtained coordinates and stored for subsequent determination of whether the flight trajectory of the bird overlaps with the operation area of ​​the wind blade.

[0097] The method further comprises:

[0098] The Kalman filter is used to estimate and predict the motion state of the target bird to improve the tracking stability and accuracy.

[0099] Through the data association algorithm, the position, size and grayscale features of the target birds in multiple frames are matched so that the detected target birds can be associated with the subsequent flight trajectory to ensure that the trajectory of each bird is independent and continuous.

[0100] It should be noted that for a more specific description of the workflow of the method embodiment, please refer to the aforementioned device embodiment part, which will not be repeated here.

[0101] The entire solution overcomes the problem that existing technologies make it difficult to trace the microscopic flight process of birds near wind turbine blades, and are unable to grasp the behavioral state of birds in response to the rotation of wind turbine blades; it solves the problem that existing assessment methods ignore the relative motion mechanism of birds and wind turbine blades, and are unable to accurately assess the substantial impact of wind blade rotation on birds; it monitors the behavioral dynamics of birds around offshore wind turbines, effectively identifies and tracks the flight trajectories of birds, and through monitoring can timely discover and grasp the behavioral changes of birds when passing through wind turbine blades, providing effective data support for subsequent scientific assessments of bird threat risks, and accurately assessing the substantial impact of wind blade rotation on birds.

[0102] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A monitoring and assessment device for the risk of bird collision with offshore wind turbine blades, characterized in that: The monitoring and evaluation device comprises: A reading module, used for reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on the wind turbine tower; A recognition module, used to extract features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result and perform labeling; wherein the recognition result includes a fan blade and a target bird; Analysis modules for: Based on the monitoring image and the recognition result, the operation area of ​​the fan blade is framed to obtain the movement trajectory of the fan blade; The specific location of the bird is marked in each frame through the preset target detection model, and connected to form the corresponding bird flight trajectory; Evaluation modules for: Based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory, determining the bird's behavior type in response to blade rotation; The corresponding proportions of the monitored birds in the behavior types are calculated, and corresponding evaluation information is obtained based on the obtained proportion data.

2. A monitoring and assessment device for the risk of bird collision with offshore wind turbine blades as claimed in claim 1, characterized in that: The identification module is further used for: The recognition result is optimized using a preset multi-layer neural network to clarify the species corresponding to the target bird.

3. The device for monitoring and evaluating the risk of bird collision with offshore wind turbine blades according to claim 1, characterized in that: The obtaining of the motion trajectory of the fan blades specifically includes: Using the cv2.rectangle function provided by OpenCV to draw a rectangular frame on the monitoring image, mark the running track of the wind blade on the monitoring image, and obtain the coordinates of the rectangular frame; The width, height and area of ​​the rectangular frame are determined based on the obtained coordinates and stored for subsequent determination of whether the flight trajectory of the bird overlaps with the operation area of ​​the wind blade.

4. A monitoring and evaluation device for the risk of bird collision with offshore wind turbine blades as claimed in claim 3, characterized in that: The analysis module is also used for: The Kalman filter is used to estimate and predict the motion state of the target bird to improve the stability and accuracy of tracking.

5. A device for monitoring and evaluating the risk of bird collision with offshore wind turbine blades according to any one of claims 1 to 4, characterized in that: The analysis module is also used for: Through the data association algorithm, the position, size and grayscale features of the target birds in multiple frames are matched so that the detected target birds can be associated with the subsequent flight trajectory to ensure that the trajectory of each bird is independent and continuous.

6. A device for monitoring and evaluating the risk of bird collision with offshore wind turbine blades as claimed in claim 5, characterized in that: The behavior types include adjusted behavior, unadjusted behavior, and collision phenomena; When a bird suddenly adjusts its direction and returns or stops flying when approaching a blade, it is considered that the bird exhibits the said adjustment behavior when the blade rotates; When the bird flies in a direction parallel to the blade, or turns multiple times within the blade's sweep area to avoid the blade, the bird is considered to exhibit the described adjustment behavior in response to the blade's rotation; When the bird enters the blade sweep area but the flight trajectory does not change, the bird is considered to exhibit the unadjusted behavior; When a bird collides with a fan blade, it is considered that the bird and the blade have the collision phenomenon.

7. A method for monitoring and assessing the risk of bird collision with offshore wind turbine blades, characterized in that: The monitoring and evaluation device for the risk of bird collision of offshore wind turbine blades as described in claim 1 comprises the following steps: Reading a monitoring image; wherein the monitoring image is collected by a monitoring device deployed on the wind turbine tower; Extracting features of the target in the monitoring image through a preset convolutional neural network to obtain a recognition result and label it; wherein the recognition result includes a fan blade and a target bird; Based on the monitoring image and the recognition result, the operation area of ​​the fan blade is framed to obtain the movement trajectory of the fan blade; The specific location of the bird is marked in each frame through the preset target detection model, and connected to form the corresponding bird flight trajectory; Based on the intersection of the bird's flight trajectory and the wind turbine blade's motion trajectory, determining the bird's behavior type in response to blade rotation; The corresponding proportions of the monitored birds in the behavior types are calculated, and corresponding evaluation information is obtained based on the obtained proportion data.

8. A method for monitoring and assessing the risk of bird collision with offshore wind turbine blades as claimed in claim 7, characterized in that: The method further comprises: The recognition result is optimized using a preset multi-layer neural network to clarify the species corresponding to the target bird.

9. A method for monitoring and assessing the risk of bird collision with offshore wind turbine blades as claimed in claim 7, characterized in that: The obtaining of the motion trajectory of the fan blades specifically includes: Using the cv2.rectangle function provided by OpenCV to draw a rectangular frame on the monitoring image, mark the running track of the wind blade on the monitoring image, and obtain the coordinates of the rectangular frame; The width, height and area of ​​the rectangular frame are determined based on the obtained coordinates and stored for subsequent determination of whether the flight trajectory of the bird overlaps with the operation area of ​​the wind blade.

10. A method for monitoring and assessing the risk of bird collision with offshore wind turbine blades according to any one of claims 7 to 9, characterized in that: The method further comprises: The Kalman filter is used to estimate and predict the motion state of the target bird to improve the tracking stability and accuracy. Through the data association algorithm, the position, size and grayscale features of the target birds in multiple frames are matched so that the detected target birds can be associated with the subsequent flight trajectory to ensure that the trajectory of each bird is independent and continuous.

Images