Intelligent bird repelling control system based on complex scene image recognition

Abstract

The invention discloses a bird repelling intelligent control system based on complex scene image recognition, and relates to the technical field of image recognition. The system comprises a complex scene bird repelling data acquisition module, a complex scene bird repelling data analysis module, a comprehensive analysis module and a bird repelling intelligent control module for optimizing and adjusting image recognition. According to the bird repelling intelligent control method, the complex scene image recognition accuracy evaluation index and the complex scene bird repelling multi-target tracking monitoring evaluation index are obtained by analyzing the complex scene bird repelling data, the complex scene bird repelling reliability evaluation index is obtained by comprehensive analysis, and the image recognition is optimized and adjusted. The accuracy of bird repelling intelligent control real-time prediction based on complex scene image recognition is improved, and the problem that the accuracy of bird repelling intelligent control real-time prediction based on complex scene image recognition is insufficient in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and in particular to an intelligent bird-repelling control system based on complex scene image recognition. Background Art

[0002] With the rapid development of civil aviation, bird strikes have become one of the important factors threatening aviation safety. Therefore, how to prevent bird strikes while protecting the ecological environment has become an important social issue. The intelligent bird repellent system can effectively reduce the occurrence of bird strikes without harming birds by using image recognition technology, big data analysis, and intelligent prediction, so as to achieve the goal of birds and aircraft sharing the blue sky. The application of the intelligent bird repellent system not only brings significant economic benefits, but also has important social benefits. By reducing bird strikes, the intelligent bird repellent system can reduce the maintenance costs of airlines and the economic losses caused by flight delays.

[0003] The existing bird-repellent intelligent control system based on complex scene image recognition uses radar detection technology to detect and identify bird targets; uses radar point cloud and visual image fusion technology, realizes radar and camera coordinate fusion through mean fitting, and uses vision and scene enhancement technology to create bird point cloud-image fusion data sets in various meteorological environments; and uses AI bird recognition algorithm to achieve real-time monitoring and intelligent recognition of bird activities.

[0004] For example, the airport bird-repellent intelligent auxiliary control system based on image recognition, which is announced in the patent application with announcement number: CN116630897B, includes: dividing birds in the flight period into suspected birds and normal birds; obtaining the flapping frequency and warning frequency of suspected birds; obtaining the credibility of suspected birds according to the difference between the flapping frequency of suspected birds and the flapping frequency of normal birds and the difference between the warning frequency of suspected birds and the warning frequency of normal birds, and obtaining untrustworthy birds among the suspected birds according to the credibility of suspected birds; obtaining the driving priority of each bird according to the number of appearances of untrustworthy birds among each bird and the total number of appearances of each bird, and performing bird driving control according to the driving priority of each bird.

[0005] For example, the patent application with publication number: CN114943984A discloses an intelligent bird detection and repelling method and device based on radar point cloud and visual image, including: according to the feature layer fusion strategy, the millimeter wave radar and camera are integrated in coordinate system transformation and neural network input format, and the front-end acquisition technology, mid-end recognition technology and back-end discrimination technology are proposed: the front-end acquisition technology can accurately realize the coordinate system fusion, the mid-end recognition technology has more advantages in accuracy and robustness than the current mainstream target detection technology, and the back-end discrimination technology can meet the actual engineering application of accurate discrimination of bird behavior, and provide a criterion for whether to start the bird repelling device; finally, a set of intelligent bird detection and repelling devices is designed by integrating the functions and relationships of the above-mentioned technologies, which provides a strong guarantee for preventing the influence of bird damage and ensuring the safe operation of transmission lines.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems: The bird-repellent intelligent control based on complex scene image recognition has the problem of insufficient real-time prediction accuracy in airport scenarios. Due to the complex airport environment, lighting changes and background clutter affect the accuracy of complex scene image recognition. At the same time, there are multiple bird targets in the airport environment, making it difficult to track and identify multiple targets, resulting in difficulty in real-time monitoring and identification of a large number of birds. The bird-repellent intelligent control based on complex scene image recognition has the problem of insufficient real-time prediction accuracy. Summary of the invention

[0007] The embodiment of the present application solves the problem of insufficient real-time prediction accuracy of bird-repellent intelligent control based on complex scene image recognition in the prior art by providing a bird-repellent intelligent control system based on complex scene image recognition, and improves the real-time prediction accuracy of bird-repellent intelligent control based on complex scene image recognition.

[0008] The embodiment of the present application provides a bird-repellent intelligent control system based on complex scene image recognition, including: a module for collecting bird-repellent data for complex scenes, a module for analyzing bird-repellent data for complex scenes, a comprehensive analysis module, and a bird-repellent intelligent control module for optimizing and adjusting image recognition: wherein the module for collecting bird-repellent data for complex scenes is used to collect and process bird-repellent data for complex scenes; the module for analyzing bird-repellent data for complex scenes is used to analyze bird-repellent data for complex scenes, obtain an evaluation index for accuracy of image recognition for complex scenes and obtain an evaluation index for multi-target tracking and monitoring of bird-repellent for complex scenes; the comprehensive analysis module is used to comprehensively analyze and obtain an evaluation index for reliability of bird-repellent for complex scenes; the bird-repellent intelligent control module for optimizing and adjusting image recognition is used to compare and analyze the evaluation index for accuracy of image recognition for complex scenes with a first threshold value of the evaluation index for accuracy of image recognition for complex scenes, and optimize and adjust the image recognition method for complex scenes; compare and analyze the evaluation index for multi-target tracking and monitoring of bird-repellent for complex scenes with a second threshold value of the evaluation index for multi-target tracking and monitoring of bird-repellent for complex scenes, and optimize and adjust the multi-target tracking and monitoring method for bird-repellent for complex scenes; compare and analyze the reliability evaluation index for complex scenes with the comprehensive threshold value of the reliability evaluation index for complex scenes, and optimize and adjust the image recognition method for bird-repellent intelligent control.

[0009] Furthermore, the specific steps of collecting and processing complex scene bird-repelling data are: collecting original complex scene bird-repelling data through cameras and sensor devices; cleaning and denoising the original complex scene bird-repelling data to obtain complex scene bird-repelling data, and the complex scene bird-repelling data includes complex scene image recognition accuracy data and complex scene bird-repelling multi-target tracking and monitoring data.

[0010] Furthermore, the specific steps for obtaining the complex scene image recognition accuracy evaluation index are: obtaining the image noise of the complex scene image recognition accuracy detection point at the first preset time through an image sensor; obtaining the image complexity entropy value of the complex scene image recognition accuracy detection point at the first preset time through image processing software; obtaining the image resolution of the complex scene image recognition accuracy detection point at the first preset time through a panoramic camera; obtaining the light intensity of the complex scene image recognition accuracy detection point at the first preset time through an automatic inspection device for light intensity of embedded lamps; obtaining the electromagnetic intensity of the complex scene image recognition accuracy detection point at the first preset time through an electromagnetic acquisition device; the complex scene image recognition accuracy data includes image noise, image complexity entropy value, image resolution, light intensity and electromagnetic intensity; and the complex scene image recognition accuracy evaluation index is obtained based on the analysis of the complex scene image recognition accuracy data.

[0011] Furthermore, the specific process of obtaining the complex scene bird-repelling multi-target tracking monitoring evaluation index is as follows: obtaining the light intensity of the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through the embedded lamp light intensity automatic inspection device; obtaining the electromagnetic intensity of the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through the electromagnetic acquisition equipment; obtaining the image noise of the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through the image sensor; obtaining the number of target birds at the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through the bird detection radar; obtaining the distance between the predicted target position and the actual target position of the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through the ultrasonic rangefinder; obtaining the image processing speed of the complex scene bird-repelling multi-target tracking monitoring detection point at the second preset time through video monitoring; the complex scene bird-repelling multi-target tracking monitoring data includes light intensity, electromagnetic intensity, image noise, the number of target birds, the distance between the predicted target position and the actual target position and the image processing speed; the complex scene bird-repelling multi-target tracking monitoring evaluation index is obtained according to the complex scene bird-repelling multi-target tracking monitoring data analysis.

[0012] Furthermore, the specific steps of the comprehensive analysis to obtain the complex scene bird repellent reliability evaluation index are: obtaining the interruption time of the complex scene bird repellent reliability detection point at the third preset time through the NTP clock synchronization device; and obtaining the complex scene bird repellent reliability evaluation index through comprehensive analysis of the complex scene image recognition accuracy evaluation index, the complex scene bird repellent multi-target tracking and monitoring evaluation index and the interruption time.

[0013] Furthermore, the specific steps of optimizing and adjusting the complex scene image recognition method are as follows: if the complex scene image recognition accuracy evaluation index is greater than or equal to the first threshold of the complex scene image recognition accuracy evaluation index, there is no need to optimize and adjust the complex scene image recognition; if the complex scene image recognition accuracy evaluation index is lower than the first threshold of the complex scene image recognition accuracy evaluation index, the aperture of the camera is controlled to adapt to different lighting changes, image denoising and filtering are performed through the hardware acceleration unit, and the GigE Vision image transmission protocol is selected.

[0014] Furthermore, the specific steps of optimizing and adjusting the complex scene bird-repelling multi-target tracking and monitoring method are as follows: if the complex scene bird-repelling multi-target tracking and monitoring evaluation index is lower than the second threshold value of the complex scene bird-repelling multi-target tracking and monitoring evaluation index, there is no need to optimize and adjust the complex scene bird-repelling multi-target tracking and monitoring method; if the complex scene bird-repelling multi-target tracking and monitoring evaluation index is greater than or equal to the second threshold value of the complex scene bird-repelling multi-target tracking and monitoring evaluation index, adjust the parameters of the tracking filter, and use the convolutional neural network to identify the characteristics of birds in various environments.

[0015] Furthermore, the specific steps of the bird-repellent intelligent control method for optimizing and adjusting image recognition are: extracting a comprehensive threshold of a complex scene bird-repellent reliability assessment index from a complex scene bird-repellent database, and comparing the complex scene bird-repellent reliability assessment index with the comprehensive threshold of the complex scene bird-repellent reliability assessment index; if the complex scene bird-repellent reliability assessment index is greater than or equal to the comprehensive threshold of the complex scene bird-repellent reliability assessment index, there is no need to optimize and adjust the image recognition bird-repellent intelligent control, and no early warning is required.

[0016] Furthermore, the method for optimizing and adjusting the image recognition-based bird-repellent intelligent control also includes: if the complex scene bird-repellent reliability evaluation index is lower than the complex scene bird-repellent reliability evaluation index comprehensive threshold, it is necessary to optimize and adjust the image recognition-based bird-repellent intelligent control, and when the target birds enter the monitoring range, an early warning is issued to notify relevant personnel to start bird repelling.

[0017] Furthermore, the bird-repellent intelligent control method for optimizing and adjusting image recognition also includes: if no warning is required, there is no need to optimize and adjust the bird-repellent intelligent control method for image recognition; if a warning is issued to notify relevant personnel to start bird repelling, Kalman filtering is used to predict the future position of the target, and edge detection is used to extract image features.

[0018] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. By analyzing the complex scene bird-repellent data, we obtain the complex scene image recognition accuracy evaluation index and the complex scene bird-repellent multi-target tracking and monitoring evaluation index. Comprehensive analysis is used to obtain the complex scene bird-repellent reliability evaluation index and optimize the image recognition-based bird-repellent intelligent control method, thereby improving the real-time prediction accuracy of the bird-repellent intelligent control based on complex scene image recognition, and solving the problem of insufficient real-time prediction accuracy of the bird-repellent intelligent control based on complex scene image recognition in the existing technology.

[0019] 2. By analyzing the bird-repelling data of complex scenes, we can understand the behavioral characteristics of birds in different environments and their adaptability in different environments. According to the behavioral characteristics of birds in different environments, we can adjust the working parameters of the bird-repelling equipment, such as sound frequency, visual stimulation intensity, etc., to improve the bird-repelling effect. At the same time, according to the adaptability of the bird-repelling system in different environments, we can choose bird-repelling equipment that is more suitable for a specific environment to improve the bird-repelling effect.

[0020] 3. By optimizing and adjusting the bird-repelling intelligent control method of image recognition, image recognition technology can monitor bird activities in real time and automatically adjust bird-repelling strategies to adapt to the behavioral characteristics of birds in different environments. By optimizing and adjusting the image recognition algorithm, the accuracy and effectiveness of bird-repelling can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of the structure of an intelligent bird-repellent control system based on complex scene image recognition provided in an embodiment of the present application; Figure 2 Schematic diagram of the complex scenario bird-repellent reliability evaluation index function provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The embodiment of the present application solves the problem of insufficient real-time prediction accuracy of bird repellent intelligent control based on complex scene image recognition in the prior art by providing a bird repellent intelligent control system based on complex scene image recognition. By analyzing complex scene bird repellent data, a complex scene image recognition accuracy evaluation index and a complex scene bird repellent multi-target tracking and monitoring evaluation index are obtained. A comprehensive analysis is performed to obtain a complex scene bird repellent reliability evaluation index and a bird repellent intelligent control method that optimizes and adjusts image recognition, thereby improving the real-time prediction accuracy of bird repellent intelligent control based on complex scene image recognition.

[0023] The technical solution in the embodiment of the present application is to solve the above-mentioned problem of insufficient accuracy of real-time prediction of bird-repelling intelligent control based on complex scene image recognition. The overall idea is as follows: By analyzing the bird-repellent data of complex scenes, we obtained the complex scene image recognition accuracy evaluation index and the complex scene bird-repellent multi-target tracking and monitoring evaluation index. The complex scene bird-repellent reliability evaluation index was obtained through comprehensive analysis and the bird-repellent intelligent control method based on image recognition was optimized and adjusted, thereby improving the real-time prediction accuracy of the bird-repellent intelligent control based on complex scene image recognition.

[0024] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0025] like Figure 1As shown, it is a structural schematic diagram of the bird-repellent intelligent control system based on complex scene image recognition provided by the embodiment of the present application. The bird-repellent intelligent control system based on complex scene image recognition provided by the embodiment of the present application includes: a module for collecting bird-repellent data for complex scenes, a module for analyzing bird-repellent data for complex scenes, a comprehensive analysis module and a bird-repellent intelligent control module for optimizing and adjusting image recognition: wherein, the module for collecting bird-repellent data for complex scenes is used to collect and process bird-repellent data for complex scenes; the module for analyzing bird-repellent data for complex scenes is used to analyze bird-repellent data for complex scenes, obtain an evaluation index for complex scene image recognition accuracy and obtain an evaluation index for multi-target tracking and monitoring of bird-repellent for complex scenes; the The comprehensive analysis module is used for comprehensive analysis to obtain the reliability evaluation index of bird-repellent in complex scenes; the bird-repellent intelligent control module for optimizing and adjusting image recognition is used for comparing and analyzing the complex scene image recognition accuracy evaluation index with the first threshold of the complex scene image recognition accuracy evaluation index, and optimizing and adjusting the complex scene image recognition method; comparing and analyzing the complex scene bird-repellent multi-target tracking and monitoring evaluation index with the second threshold of the complex scene bird-repellent multi-target tracking and monitoring evaluation index, and optimizing and adjusting the complex scene bird-repellent multi-target tracking and monitoring method; comparing and analyzing the complex scene bird-repellent reliability evaluation index with the comprehensive threshold of the complex scene bird-repellent reliability evaluation index, and optimizing and adjusting the image recognition bird-repellent intelligent control method.

[0026] Furthermore, the specific steps of collecting and processing complex scene bird-repelling data are: collecting original complex scene bird-repelling data through cameras and sensor devices; cleaning and denoising the original complex scene bird-repelling data to obtain complex scene bird-repelling data, wherein the complex scene bird-repelling data includes complex scene image recognition accuracy data and complex scene bird-repelling multi-target tracking and monitoring data.

[0027] Furthermore, the specific steps for obtaining the complex scene image recognition accuracy evaluation index are: obtaining the image noise of the complex scene image recognition accuracy detection point at the first preset time through an image sensor; obtaining the image complexity entropy value of the complex scene image recognition accuracy detection point at the first preset time through image processing software; obtaining the image resolution of the complex scene image recognition accuracy detection point at the first preset time through a panoramic camera; obtaining the light intensity of the complex scene image recognition accuracy detection point at the first preset time through an automatic inspection device for light intensity of embedded lamps; obtaining the electromagnetic intensity of the complex scene image recognition accuracy detection point at the first preset time through an electromagnetic acquisition device; the complex scene image recognition accuracy data includes image noise, image complexity entropy value, image resolution, light intensity and electromagnetic intensity; and the complex scene image recognition accuracy evaluation index is obtained based on the analysis of the complex scene image recognition accuracy data.

[0028] In this embodiment, the specific method of obtaining the complex scene image recognition accuracy evaluation index through analysis is as follows: ; ; ; ; ; The preset complex scene image recognition accuracy detection points are numbered in sequence. Indicates the number of the detection point for evaluating the accuracy of complex scene image recognition. , Indicates the total number of detection points for complex scene image recognition accuracy.

[0029] Indicates The complex scene image recognition accuracy evaluation index is a complex scene image recognition accuracy evaluation test point.

[0030] It represents the standard value of image noise, which is the preset image noise threshold obtained from the complex scene bird-repellent database. It represents the normal image noise standard value of a certain complex scene image recognition accuracy evaluation detection point. If the deviation between the image noise and the standard deviation is larger, it will affect the accuracy of image recognition in complex scenes.

[0031] Indicates The image noise under the detection point is used to evaluate the recognition accuracy of complex scene images.

[0032] It represents the standard value of image complexity entropy, which is the preset standard value of image complexity entropy obtained from the complex scene bird-repelling database. It represents the standard value of image complexity entropy of a certain complex scene image recognition accuracy evaluation detection point. The greater the deviation between the image complexity entropy value and the standard value of the image complexity entropy value, the more dispersed the information in the image is, and the greater the inaccuracy of complex scene image recognition is.

[0033] Indicates The image resolution at the detection point is used to evaluate the recognition accuracy of complex scene images.

[0034] It represents the image resolution threshold, which is the preset image resolution threshold obtained from the complex scene bird repellent database. It represents the image resolution threshold of a certain complex scene image recognition accuracy evaluation detection point. If it is lower than the image noise threshold, it will affect the accuracy of image recognition in complex scenes.

[0035] Indicates The illumination intensity correction coefficient at the detection point is used to evaluate the recognition accuracy of complex scene images.

[0036] Indicates The electromagnetic intensity correction coefficient at the detection point is used to evaluate the accuracy of complex scene image recognition.

[0037] Indicates The recognition accuracy of complex scene images is evaluated by the light intensity at the detection point.

[0038] It represents the standard value of light intensity, which is the preset standard value of light intensity obtained from the complex scene bird-repellent database. It represents the standard value of light intensity at a detection point for evaluating the accuracy of image recognition in a complex scene. It means that if the light intensity approaches the standard value of light intensity, it will not affect the accuracy of image recognition in complex scenes.

[0039] Indicates The electromagnetic intensity at the detection point is accurately evaluated by a complex scene image recognition.

[0040] It represents the standard value of electromagnetic intensity, which is the preset standard value of electromagnetic intensity obtained from the complex scene bird-repellent database. It represents the standard value of electromagnetic intensity of a certain complex scene image recognition accuracy evaluation detection point. It means that if the electromagnetic intensity approaches the standard value of electromagnetic intensity, it will not affect the accuracy of image recognition in complex scenes.

[0041] is the preset image noise impact weight factor obtained from the complex scene bird repellent database.

[0042] The weight factor affecting the preset image complexity entropy value obtained from the complex scene bird-repelling database.

[0043] It is the preset image resolution impact weight factor obtained from the complex scene bird repellent database.

[0044] It is the preset light intensity correction factor obtained from the complex scene bird repellent database.

[0045] It is the preset electromagnetic intensity correction factor obtained from the complex scene bird repellent database.

[0046] The preset image noise influence weight factor, the preset image complexity entropy value influence weight factor and the preset image resolution influence weight factor respectively represent the numerical values ​​of the influence of image noise, image complexity entropy value and image resolution on the complex scene image recognition accuracy evaluation index.

[0047] When in use, the preset image noise influence weight factor, the preset image complexity entropy value influence weight factor and the preset image resolution influence weight factor can be directly obtained from the complex scene bird-repellent database according to the image contrast value, and the corresponding relationship can be a pre-set mapping relationship. For example, a mapping set of image contrast and its corresponding preset image noise influence weight factor, preset image complexity entropy value influence weight factor and preset image resolution influence weight factor is constructed, and the real-time image contrast value is input into the mapping set to obtain the preset image noise influence weight factor, the preset image complexity entropy value influence weight factor and the preset image resolution influence weight factor. The mapping relationship can be one-to-one or many-to-one.

[0048] The preset light intensity correction factor and the preset electromagnetic intensity correction factor respectively represent the numerical values ​​of the influence of light intensity and electromagnetic intensity on the complex scene image recognition accuracy evaluation index.

[0049] When in use, the preset light intensity correction factor and the preset electromagnetic intensity correction factor can be directly obtained from the complex scene bird repellent database according to the ambient temperature value, and the corresponding relationship can be a pre-set mapping relationship. For example, a mapping set of ambient temperature and its corresponding preset light intensity correction factor and preset electromagnetic intensity correction factor is constructed, and the real-time ambient temperature value is input into the mapping set to obtain the preset light intensity correction factor and the preset electromagnetic intensity correction factor. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship.

[0050] The greater the deviation between the light intensity and the standard value of the light intensity, the increase in image noise, the decrease in the sensor's sensitivity to light, and the decrease in the signal-to-noise ratio. The light intensity affects the contrast and details of the image, thereby affecting the complexity entropy value of the image. When the light intensity approaches the standard value of the light intensity, the image complexity entropy value can be increased, and the richness of information can be increased. Electromagnetic interference will increase image noise. The electromagnetic interference signal is captured by the sensor and appears as noise in the image. Electromagnetic interference reduces the entropy value of the image. Electromagnetic interference causes irregular noise patterns in the image, reducing the amount of information in the image. Image noise reduces the resolution of the image. Image noise blurs the image details, making the image look more blurred.

[0051] The smaller the square of the difference between the illumination intensity and the standard value of illumination intensity, the higher the contrast and clarity of the image, which is negatively correlated with the complex scene image recognition accuracy evaluation index, and the higher the complex scene image recognition accuracy evaluation index; the smaller the square of the difference between the electromagnetic intensity and the standard value of electromagnetic intensity, the reduced image noise, which is negatively correlated with the complex scene image recognition accuracy evaluation index, and the higher the complex scene image recognition accuracy evaluation index; the larger the absolute value of the difference between the image noise and the standard value of image noise, the interference with the extraction of recognition image features, which is negatively correlated with the complex scene image recognition accuracy evaluation index, and the lower the complex scene image recognition accuracy evaluation index; the image complexity entropy value reflects the diversity of image content, and the lower the absolute value of the difference between the image complexity entropy value and the standard value of the image complexity entropy value, the more information the image contains, which is negatively correlated with the complex scene image recognition accuracy evaluation index, and the higher the complex scene image recognition accuracy evaluation index; the higher the image resolution, the more image details it contains, which helps to identify more image features, and is positively correlated with the complex scene image recognition accuracy evaluation index, and the higher the complex scene image recognition accuracy evaluation index.

[0052] Furthermore, the specific process of obtaining the evaluation index of bird-repelling multi-target tracking and monitoring in complex scenes is as follows: obtain the light intensity of the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through the embedded lamp light intensity automatic inspection device; obtain the electromagnetic intensity of the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through the electromagnetic acquisition equipment; obtain the image noise of the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through the image sensor; obtain the number of target birds at the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through the bird detection radar; obtain the distance between the predicted target position and the actual target position of the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through the ultrasonic rangefinder; obtain the image processing speed of the bird-repelling multi-target tracking and monitoring detection point in complex scenes at the second preset time through video monitoring; the complex scene bird-repelling multi-target tracking and monitoring data include light intensity, electromagnetic intensity, image noise, the number of target birds, the distance between the predicted target position and the actual target position and the image processing speed; the bird-repelling multi-target tracking and monitoring evaluation index in complex scenes is obtained according to the analysis of the bird-repelling multi-target tracking and monitoring data in complex scenes.

[0053] In this embodiment, the specific method for analyzing and obtaining the evaluation index of bird-repelling multi-target tracking monitoring in complex scenes is as follows: ; ; ; ; ; ; ; The preset complex scene bird-repelling multi-target tracking monitoring detection points are numbered in sequence. Indicates The number of the complex scene bird-repelling multi-target tracking monitoring detection point under the complex scene bird-repelling multi-target tracking monitoring detection segment, , S represents the total number of detection points for bird-repelling multi-target tracking and monitoring in complex scenes.

[0054] Divide the preset complex scene bird-repelling multi-target tracking monitoring time into complex scene bird-repelling multi-target tracking monitoring detection segments of the same length. Indicates the number of the complex scene bird-repelling multi-target tracking monitoring detection segment. , Indicates the total number of detection segments for bird-repelling multi-target tracking and monitoring in complex scenarios.

[0055] Indicates The complex scene bird-repellent multi-target tracking monitoring evaluation index of the complex scene bird-repellent multi-target tracking monitoring detection points.

[0056] Indicates Light intensity correction coefficient for a complex scene bird-repelling multi-target tracking monitoring detection point.

[0057] Indicates Electromagnetic intensity correction coefficient of a complex scene bird-repelling multi-target tracking monitoring detection point.

[0058] Indicates Image noise correction coefficient for multiple bird-repelling multi-target tracking monitoring detection points in complex scenes.

[0059] Indicates The number of target birds in a complex scene bird-repelling multi-target tracking monitoring detection segment.

[0060] It indicates the threshold of the number of monitored target birds, which is the preset threshold of the number of monitored target birds obtained from the complex scene bird-repelling database. It indicates the upper limit of the maximum number of monitored target birds. When the number of target birds approaches the threshold of the number of target birds, the more targets are tracked, the higher the tracking difficulty.

[0061] Indicates The two-dimensional spatial distance error coefficient of the detection point under the bird-repelling multi-target tracking monitoring in a complex scene represents the distance between the predicted target position and the target position.

[0062] Indicates The image processing speed of multiple target tracking monitoring detection points in a complex scene of bird repelling is improved.

[0063] Represents the image processing speed threshold, which is the preset image processing speed threshold obtained from the complex scene bird repellent database.

[0064] Indicates The light intensity at the detection point is monitored by multi-target tracking of bird-repelling in a complex scene.

[0065] It represents the standard value of light intensity, which is the preset standard value of light intensity obtained from the complex scene bird repellent database. It represents the standard value of light intensity at a certain complex scene bird repellent multi-target tracking monitoring detection point. The greater the deviation between the light intensity and the standard value of light intensity, the greater the difficulty of complex scene bird repellent multi-target tracking.

[0066] Indicates The electromagnetic intensity of the bird-repellent multi-target tracking monitoring point in a complex scene.

[0067] It represents the standard value of electromagnetic intensity, which is the preset standard value of electromagnetic intensity obtained from the complex scene bird repellent database. It represents the standard value of electromagnetic intensity of a complex scene bird repellent multi-target tracking monitoring detection point. The greater the deviation between the electromagnetic intensity and the standard value of electromagnetic intensity, the greater the difficulty of complex scene bird repellent multi-target tracking.

[0068] Indicates Image noise at the detection point of bird-repelling multi-target tracking monitoring in a complex scene.

[0069] It represents the standard value of image noise, which is the preset standard value of image noise obtained from the complex scene bird repellent database. It represents the standard value of image noise of a complex scene bird repellent multi-target tracking monitoring detection point. If the image noise is greater than the standard value of image noise, the complex scene bird repellent multi-target tracking is more difficult.

[0070] Indicates The lateral target bird distance in two-dimensional space under a complex scene bird-repelling multi-target tracking monitoring detection point.

[0071] It represents the standard value of the lateral target bird distance, which is the preset standard value of the lateral target bird distance obtained from the complex scene bird repellent database. It represents the standard value of the lateral target bird distance of a complex scene bird repellent multi-target tracking and monitoring detection point. If the lateral target bird distance is greater than the standard value of the lateral target bird distance, the more difficult the complex scene bird repellent multi-target tracking is.

[0072] Indicates The vertical target bird distance in two-dimensional space under a complex scene bird-repelling multi-target tracking monitoring detection point.

[0073] It represents the standard value of vertical target bird distance, which is the preset standard value of vertical target bird distance obtained from the complex scene bird repellent database. It represents the standard value of vertical target bird distance of a complex scene bird repellent multi-target tracking and monitoring detection point. If the vertical target bird distance is greater than the standard value of vertical target bird distance, the complex scene bird repellent multi-target tracking is more difficult.

[0074] The weight factor is the preset target bird number obtained from the complex scene bird-repelling database.

[0075] The preset distance error coefficient influencing weight factor obtained from the complex scene bird-repelling database.

[0076] The weight factor affecting the processing speed of preset images obtained from the complex scene bird-repelling database.

[0077] It is the preset light intensity correction factor obtained from the complex scene bird repellent database.

[0078] It is the preset electromagnetic intensity correction factor obtained from the complex scene bird repellent database.

[0079] is the preset image noise correction factor obtained from the complex scene bird repellent database.

[0080] The preset target bird number influence weight factor, the preset distance error coefficient influence weight factor and the preset image processing speed influence weight factor respectively represent the numerical values ​​of the influence of the target bird number, the distance error coefficient and the image processing speed on the complex scene bird-repelling multi-target tracking monitoring evaluation index.

[0081] When in use, the preset target bird number influence weight factor, the preset distance error coefficient influence weight factor and the preset image processing speed influence weight factor can be directly obtained from the complex scene bird-repellent database according to the value of weather changes, and the corresponding relationship can be a pre-set mapping relationship. For example, a mapping set of weather changes and their corresponding preset target bird number influence weight factors, preset distance error coefficient influence weight factors and preset image processing speed influence weight factors is constructed, and the real-time weather change value is input into the mapping set to obtain the preset target bird number influence weight factor, the preset distance error coefficient influence weight factor and the preset image processing speed influence weight factor, wherein the mapping relationship can be one-to-one or many-to-one.

[0082] The preset light intensity correction factor, the preset electromagnetic intensity correction factor and the preset image noise correction factor respectively represent the numerical values ​​of the influence of light intensity, electromagnetic intensity and image noise on the complex scene bird-repelling multi-target tracking and monitoring evaluation index.

[0083] When in use, the preset light intensity correction factor, preset electromagnetic intensity correction factor and preset image noise correction factor can be directly obtained from the complex scene bird repellent database according to the value of the target bird number, and the corresponding relationship can be a pre-set mapping relationship. For example, a mapping set of the target bird number and its corresponding preset light intensity correction factor, preset electromagnetic intensity correction factor and preset image noise correction factor is constructed, and the real-time target bird number value is input into the mapping set to obtain the preset light intensity correction factor, preset electromagnetic intensity correction factor and preset image noise correction factor. The mapping relationship can be a one-to-one correspondence or a many-to-one relationship.

[0084] The greater the deviation between the light intensity and the standard value of the light intensity, the increase in image noise, because the sensor's sensitivity to light decreases, resulting in a decrease in the signal-to-noise ratio. The higher the electromagnetic intensity, the increase in image noise, affecting the image quality. The increase in image noise increases the complexity of image processing, thereby reducing the image processing speed. The increase in the number of target birds increases the amount of processed image data and reduces the image processing speed. The increase in the number of target birds will increase the difficulty of tracking, resulting in an increase in the distance error coefficient. The greater the distance between the predicted bird position and the actual bird position, the lower the image processing speed.

[0085] The greater the square of the difference between the light intensity and the standard value of the light intensity, the lower the image quality, the more difficult it is to identify and track birds, and there is a positive correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index. The larger the complex scene bird-repelling multi-target tracking monitoring and evaluation index is; the greater the square of the difference between the electromagnetic intensity and the standard value of the electromagnetic intensity, the increase in image noise and the lower the image quality, and there is a positive correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index. The larger the complex scene bird-repelling multi-target tracking monitoring and evaluation index is; the greater the square of the difference between the image noise and the standard value of the image noise, the more difficult it is to identify and track birds, and there is a positive correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index. The larger the target tracking monitoring and evaluation index is; the larger the number of target birds is, the greater the complexity of tracking is, resulting in a decrease in tracking performance, and there is a positive correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index, and the larger the complex scene bird-repelling multi-target tracking monitoring and evaluation index is; an increase in the distance error coefficient means that the gap between the predicted bird position and the actual position increases, and there is a positive correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index, and the larger the complex scene bird-repelling multi-target tracking monitoring and evaluation index is; the higher the image processing speed, the faster the image data processing is, there is a negative correlation with the complex scene bird-repelling multi-target tracking monitoring and evaluation index, and the lower the complex scene bird-repelling multi-target tracking monitoring and evaluation index is.

[0086] Furthermore, the specific steps of comprehensively analyzing and obtaining the complex scene bird repellent reliability evaluation index are: obtaining the interruption time of the complex scene bird repellent reliability detection point at the third preset time through the NTP clock synchronization device; and obtaining the complex scene bird repellent reliability evaluation index through comprehensive analysis of the complex scene image recognition accuracy evaluation index, the complex scene bird repellent multi-target tracking and monitoring evaluation index and the interruption time.

[0087] In this embodiment, the specific method for analyzing and obtaining the reliability evaluation index of bird repellent in complex scenarios is as follows: ; ; The preset complex scene bird-repellent reliability detection segments are numbered in sequence. Indicates the number of the reliability detection segment of bird repellent in complex scenarios. , Indicates the total number of segments for detecting bird repellent reliability in complex scenarios.

[0088] Represents the reliability evaluation index of bird repellent in complex scenarios.

[0089] Indicates The complex scene image recognition accuracy evaluation index is a complex scene image recognition accuracy evaluation test point.

[0090] Indicates The complex scene bird-repellent multi-target tracking monitoring evaluation index of the complex scene bird-repellent multi-target tracking monitoring detection points.

[0091] Indicates The interruption time under the bird-repellent reliability detection segment of a complex scenario is to track the start and end time points of the interruption and the duration of each interruption.

[0092] It represents the interruption time threshold, which is the preset interruption time threshold obtained from the complex scene bird repellent database. It represents the interruption time threshold under a complex scene bird repellent reliability detection segment. If the interruption time is greater than the interruption time threshold, the reliability of the complex scene bird repellent is lower.

[0093] The weight factor affecting the recognition accuracy evaluation index of the preset complex scene images obtained from the complex scene bird repellent database.

[0094] It is the influencing weight factor of the preset complex scene bird-repelling multi-target tracking and monitoring evaluation index obtained from the complex scene bird-repelling database.

[0095] It is the preset interruption time impact weight factor obtained from the complex scene bird repellent database.

[0096] The preset complex scene image recognition accuracy evaluation index influencing weight factor, the preset complex scene bird repellent multi-target tracking and monitoring evaluation index influencing weight factor and the preset interruption time influencing weight factor respectively represent the numerical values ​​of the influence of the complex scene image recognition accuracy evaluation index, the complex scene bird repellent multi-target tracking and monitoring evaluation index and the interruption time on the complex scene bird repellent reliability evaluation index.

[0097] When in use, the preset complex scene image recognition accuracy evaluation index influencing weight factor, the preset complex scene bird repellent multi-target tracking and monitoring evaluation index influencing weight factor and the preset interruption time influencing weight factor can be directly obtained from the complex scene bird repellent database according to the value of the bird speed. The corresponding relationship can be a pre-set mapping relationship. For example, a mapping set of the bird speed and its corresponding preset complex scene image recognition accuracy evaluation index influencing weight factor, the preset complex scene bird repellent multi-target tracking and monitoring evaluation index influencing weight factor and the preset interruption time influencing weight factor is constructed, and the real-time bird speed value is input into the mapping set to obtain the preset complex scene image recognition accuracy evaluation index influencing weight factor, the preset complex scene bird repellent multi-target tracking and monitoring evaluation index influencing weight factor and the preset interruption time influencing weight factor. The mapping relationship can be one-to-one or many-to-one.

[0098] Table 1 is an example table of complex scene bird-repellent reliability evaluation index. The example parameters in Table 1 only take the parameters of a complex scene bird-repellent reliability detection point for example. Set to 0.4, weight factor Set to 0.3, weight factor Set to 0.3, The interruption time threshold is 2, as shown in Table 1.

[0099] Table 1 Example of reliability evaluation index of bird repellent in complex scenarios Complex scene image recognition accuracy evaluation index Evaluation index of bird-repelling multi-target tracking monitoring in complex scenes Interruption time Reliability evaluation index of bird repellent in complex scenarios 2 1 3 3.22061762305 3 2 2 2.02700256506 1 2 1 1.41005286453 4 4 0 1.74870667749 6 5 5 0.973544714305 From Table 1 and Figure 2 It can be obtained that the complex scene image recognition accuracy evaluation index is the basis of complex scene target tracking. If the complex scene image recognition is inaccurate, multi-target tracking will be affected. The higher the complex scene image recognition accuracy, the higher the accuracy of complex scene bird-repelling multi-target tracking. The image recognition accuracy is affected by the interruption time. The longer the interruption time, the lower the image recognition accuracy, and the new image data cannot be processed in time. The longer the interruption time, the loss of tracking of complex scene targets, and the complex scene targets will move to new positions in a short time.

[0100] The higher the image recognition accuracy, the more accurately birds and other targets can be identified, thereby improving the reliability of bird repelling. It is positively correlated with the reliability evaluation index of bird repelling in complex scenes. The higher the image recognition accuracy evaluation index of complex scenes, the higher the reliability evaluation index of bird repelling in complex scenes; the higher the multi-target tracking and monitoring evaluation index of bird repelling in complex scenes, the inability to effectively track multiple targets, resulting in a decrease in the reliability of bird repelling. It is negatively correlated with the reliability evaluation index of bird repelling in complex scenes. The higher the multi-target tracking and monitoring evaluation index of bird repelling in complex scenes, the lower the reliability evaluation index of bird repelling in complex scenes; the longer the interruption time, the inability to effectively identify and track birds, reducing the reliability of bird repelling. The longer the interruption time leads to the inability to track the target and respond to new threats in time. It is negatively correlated with the reliability evaluation index of bird repelling in complex scenes. The longer the interruption time, the lower the reliability evaluation index of bird repelling in complex scenes.

[0101] like Figure 2 As shown, it is a schematic diagram of the complex scene bird-repelling reliability evaluation index function provided in an embodiment of the present application; x is the positive half axis of the horizontal coordinate, and y is the positive half axis of the vertical coordinate.

[0102] Curve a indicates that if the complex scene bird-repellent multi-target tracking and monitoring evaluation index is set to a fixed value of 1, the interruption time is set to a fixed value of 1, the complex scene image recognition accuracy evaluation index is x, and the complex scene bird-repellent reliability evaluation index is y, the complex scene bird-repellent reliability evaluation index increases with the increase of the complex scene image recognition accuracy evaluation index.

[0103] Furthermore, the specific steps of optimizing and adjusting the complex scene image recognition method are as follows: if the complex scene image recognition accuracy evaluation index is greater than or equal to the first threshold of the complex scene image recognition accuracy evaluation index, there is no need to optimize and adjust the complex scene image recognition; if the complex scene image recognition accuracy evaluation index is lower than the first threshold of the complex scene image recognition accuracy evaluation index, the aperture of the camera is controlled to adapt to different lighting changes, image denoising and filtering are performed through the hardware acceleration unit, and the GigE Vision image transmission protocol is selected.

[0104] In this embodiment, it is assumed that the complex scene image recognition accuracy evaluation index is 85%. If the first threshold of the complex scene image recognition accuracy evaluation index is 90%. Since 85% is lower than 90%, it is necessary to adjust the aperture of the camera, use the hardware acceleration unit for image processing, and select the GigE Vision image transmission protocol. Assume that the original aperture is set to f / 4.0. In order to adapt to different lighting changes, it is adjusted to f / 2.8 to allow more light to enter, which helps to improve image quality under low light conditions. Assume that the processing time of image denoising and filtering is originally 100ms. Through the hardware acceleration unit, this time can be reduced to 50ms. Assume that the original data transmission delay is 200ms. After using the GigE Vision protocol, the delay is reduced to 100ms, and the aperture of the camera is adjusted from f / 4.0 to f / 2.8, increasing the aperture value of 1.2. The image processing time is reduced from 100 milliseconds to 50 milliseconds, a reduction of 50 milliseconds, and the data transmission delay is reduced from 200 milliseconds to 100 milliseconds, a reduction of 100 milliseconds.

[0105] Furthermore, the specific steps for optimizing and adjusting the complex scene bird-repelling multi-target tracking and monitoring method are as follows: if the complex scene bird-repelling multi-target tracking and monitoring evaluation index is lower than the second threshold value of the complex scene bird-repelling multi-target tracking and monitoring evaluation index, there is no need to optimize and adjust the complex scene bird-repelling multi-target tracking and monitoring method; if the complex scene bird-repelling multi-target tracking and monitoring evaluation index is greater than or equal to the second threshold value of the complex scene bird-repelling multi-target tracking and monitoring evaluation index, adjust the parameters of the tracking filter, and use the convolutional neural network to identify the characteristics of birds in various environments.

[0106] In this embodiment, it is assumed that the evaluation index of the bird-repelling multi-target tracking monitoring and evaluation index in complex scenes is 95%. The second threshold value of the bird-repelling multi-target tracking monitoring and evaluation index in complex scenes is 90%. Since 95% is greater than 90%, it is necessary to adjust the parameters of the tracking filter and use a convolutional neural network to identify bird features in various environments. It is assumed that the original filter parameters are set to prioritize tracking speed. In order to improve accuracy, the parameters are adjusted to prioritize tracking accuracy. It is assumed that the accuracy of bird feature recognition is 80% without the use of a convolutional neural network. By using a convolutional neural network, this accuracy can be increased to 95%.

[0107] Furthermore, the specific steps of optimizing and adjusting the image recognition bird repellent intelligent control method are as follows: extracting the comprehensive threshold of the complex scene bird repellent reliability assessment index from the complex scene bird repellent database, and comparing the complex scene bird repellent reliability assessment index with the comprehensive threshold of the complex scene bird repellent reliability assessment index; if the complex scene bird repellent reliability assessment index is greater than or equal to the comprehensive threshold of the complex scene bird repellent reliability assessment index, there is no need to optimize and adjust the image recognition bird repellent intelligent control, and no early warning is required.

[0108] Furthermore, the bird-repellent intelligent control method for optimizing and adjusting image recognition also includes: if the complex scene bird-repellent reliability evaluation index is lower than the complex scene bird-repellent reliability evaluation index comprehensive threshold, it is necessary to optimize and adjust the image recognition bird-repellent intelligent control, and when the target birds enter the monitoring range, an early warning is issued to notify relevant personnel to start bird repelling.

[0109] In this embodiment, first, the system detects and identifies the target birds entering the monitoring range through the camera and image processing technology, calculates the reliability evaluation index of bird repellent in complex scenes, and determines whether it is lower than the preset comprehensive threshold. If the reliability evaluation index of bird repellent in complex scenes is lower than the threshold, the system will trigger the early warning mechanism and send early warning information to relevant personnel through preset communication methods, such as SMS, email, application push notification, etc. The early warning information should include the type, number, location and possible threat level of the birds. The relevant personnel who receive the early warning notification should take immediate action and select appropriate bird repellent measures according to the evaluation of the early warning information, such as using sound bird repellent, visual bird repellent or other physical bird repellent methods. During the early warning response process, the system should continuously monitor bird activities and update the early warning information as needed. When the target birds leave the monitoring range or the threat is lifted, the system should issue an early warning release notification to inform the relevant personnel.

[0110] Furthermore, the bird-repellent intelligent control method for optimizing and adjusting image recognition also includes: if there is no need to issue an early warning, there is no need to optimize and adjust the bird-repellent intelligent control method for image recognition; if an early warning is issued to notify relevant personnel to start bird repelling, Kalman filtering is used to predict the future position of the target, and edge detection is used to extract image features.

[0111] In this embodiment, it is assumed that there is a complex scene bird-repelling system. When an early warning is issued to notify relevant personnel to start bird-repelling, it is necessary to use Kalman filtering to predict the future position of the target bird, and use edge detection to extract image features. Assume that the current position of the bird is (x=100, y=200) and the speed is (vx=5, vy=3). Use Kalman filtering to predict the position at the next moment. Assuming there is an image, an edge detection algorithm such as Canny edge detection is used to extract edge features in the image. According to the prediction of Kalman filtering, the position of the bird at the next moment will be (x=105, y=203). Edge detection is an image processing technology used to detect edges in an image, that is, the gradient between each pixel and its surrounding pixels. Commonly used edge detection algorithms include Canny edge detection, Sobel edge detection, etc. These algorithms detect edges by calculating the gradient between each pixel in the image and its surrounding pixels. For example, if there is an image, after using the Canny edge detection algorithm, the edge features of the image can be obtained, which can help better identify and track birds. Since this process involves image processing, it usually requires the use of a specialized image processing library (such as Open CV) to implement it. Relevant personnel can use airport-specific ultrasonic bird repellers to repel birds.

[0112] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0113] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0114] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0115] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0116] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0117] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A bird-repellent intelligent control system based on complex scene image recognition, characterized in that: It includes a module for collecting bird-repelling data in complex scenarios, a module for analyzing bird-repelling data in complex scenarios, a comprehensive analysis module, and a bird-repelling intelligent control module for optimizing and adjusting image recognition: Wherein, the module for collecting bird-repelling data for complex scenes is used to collect and process bird-repelling data for complex scenes; The complex scene bird-repelling data analysis module is used to analyze the complex scene bird-repelling data to obtain a complex scene image recognition accuracy evaluation index and a complex scene bird-repelling multi-target tracking monitoring evaluation index; The comprehensive analysis module is used to comprehensively analyze and obtain a bird-repelling reliability evaluation index for complex scenarios; The bird-repellent intelligent control module for optimizing and adjusting image recognition is used to compare and analyze the complex scene image recognition accuracy evaluation index with the first threshold of the complex scene image recognition accuracy evaluation index, and optimize and adjust the complex scene image recognition method; compare and analyze the complex scene bird-repellent multi-target tracking and monitoring evaluation index with the second threshold of the complex scene bird-repellent multi-target tracking and monitoring evaluation index, and optimize and adjust the complex scene bird-repellent multi-target tracking and monitoring method; compare and analyze the complex scene bird-repellent reliability evaluation index with the comprehensive threshold of the complex scene bird-repellent reliability evaluation index, and optimize and adjust the image recognition bird-repellent intelligent control method.

2. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The specific steps of collecting and processing complex scene bird-repelling data are as follows: Collect original bird-repelling data in complex scenes through cameras and sensor devices; The original bird-repelling data for complex scenes are cleaned and denoised to obtain the bird-repelling data for complex scenes, wherein the bird-repelling data for complex scenes include the accuracy data of image recognition for complex scenes and the multi-target tracking and monitoring data of bird-repelling for complex scenes.

3. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 2, characterized in that: The specific steps of obtaining the complex scene image recognition accuracy evaluation index are: Obtaining image noise at a first preset time complex scene image recognition accuracy detection point through an image sensor; Obtaining the image complexity entropy value of the complex scene image recognition accuracy detection point at a first preset time by using image processing software; Obtaining, by means of a panoramic camera, an image resolution of a first preset time complex scene image recognition accuracy detection point; The light intensity of the complex scene image recognition accuracy detection point at the first preset time is obtained by the embedded lamp light intensity automatic inspection device; Obtaining the electromagnetic intensity of a complex scene image recognition accuracy detection point at a first preset time through an electromagnetic acquisition device; The complex scene image recognition accuracy data includes image noise, image complexity entropy value, image resolution, light intensity and electromagnetic intensity; The complex scene image recognition accuracy evaluation index is obtained based on the complex scene image recognition accuracy data analysis.

4. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 2, characterized in that: The specific process of obtaining the complex scene bird-repelling multi-target tracking monitoring evaluation index is as follows: The light intensity of the second preset time complex scene bird-repelling multi-target tracking monitoring detection point is obtained through the embedded lamp light intensity automatic inspection device; The electromagnetic intensity of the bird-repelling multi-target tracking monitoring detection point in a complex scene at a second preset time is obtained through an electromagnetic acquisition device; The image noise of the detection point of the bird-repelling multi-target tracking monitoring in a complex scene at a second preset time is obtained by an image sensor; The number of target birds at the second preset time complex scene bird-repelling multi-target tracking monitoring detection point is obtained through the bird-detecting radar; The distance between the predicted target position and the actual target position of the bird-repelling multi-target tracking monitoring detection point in a complex scene at a second preset time is obtained by an ultrasonic rangefinder; The image processing speed of the second preset time complex scene bird-repelling multi-target tracking monitoring detection point is obtained through video monitoring; The complex scene bird-repelling multi-target tracking monitoring data includes light intensity, electromagnetic intensity, image noise, number of target birds, distance between predicted target position and actual target position, and image processing speed; The complex scene bird-repellent multi-target tracking monitoring evaluation index was obtained based on the complex scene bird-repellent multi-target tracking monitoring data analysis.

5. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The specific steps of obtaining the complex scene bird-repelling reliability evaluation index through comprehensive analysis are as follows: Obtain the interruption time of the reliability detection point of bird-repelling in complex scenarios at a third preset time through the NTP clock synchronization device; The complex scene bird repellent reliability evaluation index is obtained through comprehensive analysis of the complex scene image recognition accuracy evaluation index, the complex scene bird repellent multi-target tracking and monitoring evaluation index and the interruption time.

6. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The specific steps of optimizing and adjusting the complex scene image recognition method are as follows: If the complex scene image recognition accuracy evaluation index is greater than or equal to the complex scene image recognition accuracy evaluation index first threshold, then there is no need to optimize and adjust the complex scene image recognition; If the complex scene image recognition accuracy evaluation index is lower than the first threshold of the complex scene image recognition accuracy evaluation index, the aperture of the camera is controlled to adapt to different lighting changes, image denoising and filtering are performed through the hardware acceleration unit, and the GigE Vision image transmission protocol is selected.

7. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The specific steps of optimizing and adjusting the complex scene bird-repelling multi-target tracking and monitoring method are as follows: If the complex scene bird-repelling multi-target tracking monitoring evaluation index is lower than the complex scene bird-repelling multi-target tracking monitoring evaluation index second threshold value, there is no need to optimize and adjust the complex scene bird-repelling multi-target tracking monitoring method; If the complex scene bird-repelling multi-target tracking monitoring evaluation index is greater than or equal to the second threshold of the complex scene bird-repelling multi-target tracking monitoring evaluation index, the parameters of the tracking filter are adjusted and the convolutional neural network is used to identify the characteristics of birds in various environments.

8. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The specific steps of the bird-repelling intelligent control method for optimizing and adjusting image recognition are as follows: Extracting the comprehensive threshold of the reliability evaluation index of bird repellent in complex scenes from the bird repellent database of complex scenes, and comparing the reliability evaluation index of bird repellent in complex scenes with the comprehensive threshold of the reliability evaluation index of bird repellent in complex scenes; If the complex scene bird repellent reliability assessment index is greater than or equal to the complex scene bird repellent reliability assessment index comprehensive threshold, there is no need to optimize and adjust the image recognition bird repellent intelligent control, and there is no need to issue an early warning.

9. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The bird-repelling intelligent control method for optimizing and adjusting image recognition also includes: If the reliability assessment index of bird repellent in complex scenes is lower than the comprehensive threshold of the reliability assessment index of bird repellent in complex scenes, it is necessary to optimize and adjust the bird repellent intelligent control of image recognition. When the target birds enter the monitoring range, an early warning will be issued to notify relevant personnel to start bird repellent.

10. The bird-repelling intelligent control system based on complex scene image recognition as claimed in claim 1, characterized in that: The bird-repelling intelligent control method for optimizing and adjusting image recognition also includes: If there is no need to issue an early warning, there is no need to optimize and adjust the bird-repelling intelligent control method based on image recognition; If an early warning is issued to notify relevant personnel to start bird repelling, Kalman filtering is used to predict the future position of the target and edge detection is used to extract image features.

Images