CN-121982623-A - Fine classification and risk prevention and control method, control center and system for fusion detection of birds at airport

CN121982623ACN 121982623 ACN121982623 ACN 121982623ACN-121982623-A

Abstract

The embodiment of the invention discloses a method and a system for fine classification and risk prevention and control of fusion detection of birds at an airport; the method comprises the steps of obtaining ultra-low-altitude bird activity monitoring data through a collaborative detection system, carrying out fusion treatment on the ultra-low-altitude bird activity monitoring data, classifying according to bird species and activity modes to obtain fine classification results, and carrying out risk prediction and prevention and control according to the fine classification results. The invention has the advantages that 1, the invention designs a system for cooperatively detecting the radar and various optical detection devices, improves the activity monitoring capability of the ultra-low altitude birds, and fills the short message board. 2. The new bird classification mode is combed, expert knowledge is combined with application data on the basis, and a classical method is combined with a deep learning method, so that a classification method which can adapt to various conditions is provided, and the classification result is reliable and easy to use. 3. The fine classification information is combined with the risk prevention and control method, so that the accuracy of risk prediction is improved, and the risk prevention and control effect is improved.

Inventors

WU HONGGANG
HE DONGLIN
SUI YUNFENG
LI YANLING
JIANG YUE

Assignees

中国民用航空总局第二研究所

Dates

Publication Date: 20260505
Application Date: 20251211

Claims (9)

1. An airport bird fusion detection fine classification and risk prevention and control method is characterized by comprising the following steps: The method comprises the steps of acquiring ultra-low altitude bird activity monitoring data through a collaborative detection system, wherein the collaborative detection system comprises radar detection equipment, optical detection equipment and tracking shooting equipment; carrying out fusion processing on the ultralow-altitude bird activity monitoring data, and classifying according to bird species and activity modes to obtain a fine classification result; and carrying out risk prediction and prevention and control according to the fine classification result.
2. The method of claim 1, wherein ultra-low altitude bird activity monitoring data is obtained by a collaborative detection system, in particular: The radar detection equipment is adopted to perform large-range and ultra-low-altitude bird activity detection to obtain radar detection data; Detecting an ultra-low altitude bird target by adopting optical detection equipment to obtain optical detection data; And guiding the tracking shooting equipment to search for and lock a suspected tracking target according to the radar detection data and the optical detection data to obtain tracking shooting data.
3. The method according to claim 1, wherein the fine classification result is obtained specifically as: The ultra-low altitude bird activity monitoring data is processed by adopting a fusion processing network to obtain classification characteristics, wherein the classification characteristics comprise fusion bird characteristics, micro Doppler characteristics and image characteristics; inputting the classification features into a hierarchical classifier to obtain a fusion classification result; and (5) calling a professional knowledge base to correct the fusion classification result to obtain a fine classification result.
4. A method according to claim 3, characterized in that the classification features are obtained in particular as: extracting track features from the radar detection data and the optical detection data by adopting the track processing module, inputting the track features into a track classifier, and calculating a track type; Extracting first features from the radar detection data and extracting second features from the radar detection data and the tracking shooting data by adopting the bird feature processing module, wherein the first features comprise target size and population quantity, and the second features comprise target size, population quantity, vibration wing frequency and contour morphology; adopting the bird feature processing module to perform fusion processing on the track features, the track types, the first features, the second features and the environmental features to generate fusion bird features; performing feature coding on micro Doppler data in the radar detection data by adopting a micro Doppler information coding module, and outputting micro Doppler features; and adopting an image information coding module to perform feature coding on a plurality of images in the optical detection data and the tracking shooting data, and outputting image features.
5. The method according to claim 4, wherein extracting track features is specifically: Combining the radar detection data with the moving track in the optical detection data; preprocessing the combined moving tracks, including abnormal point removal, missing point interpolation and smoothing; And extracting speed, acceleration, altitude, direction, statistical characteristics, duration and total distance from the preprocessed moving track to form track characteristics.
6. The method of claim 3, wherein the hierarchical classifier comprises a bird feature classification module, a micro-doppler feature classification module, and an image feature classification module, and the fused classification result is specifically: adopting three classification modules to respectively process the fusion bird features, the micro Doppler features and the image features, and outputting three primary classification results; and carrying out weighted fusion and rationality treatment on the three primary classification results according to the weights set by the three classification modules to obtain a fusion classification result.
7. The method of claim 6, wherein the method further comprises: initializing the bird feature classification module by adopting manual experience, wherein only a small amount of error data is manually marked in the operation process of the hierarchical classifier; And training the micro Doppler feature classification module and the image feature classification module by adopting the results obtained by the bird feature module.
8. A control center comprising a processor, an input device, an output device and a memory, the processor, the input device, the output device and the memory being interconnected, characterized in that the memory is adapted to store a computer program comprising program instructions, the processor being configured to invoke the program instructions to perform the method according to any of claims 1-7.
9. An airport bird fusion detection fine classification and risk prevention and control system comprises radar detection equipment, optical detection equipment, tracking shooting equipment, a control center and bird repelling equipment, wherein the control center is characterized in that the control center receives data acquired by the radar detection equipment, the optical detection equipment and the tracking shooting equipment and controls the bird repelling equipment to execute a bird repelling strategy.

Description

Fine classification and risk prevention and control method, control center and system for fusion detection of birds at airport Technical Field The invention relates to the technical field of analysis of active birds in an airport flight area, in particular to a method, a control center and a system for fusion detection fine classification and risk prevention and control of the birds in the airport. Background Monitoring airport and peripheral bird activities, identifying high-risk bird activities and taking effective risk prevention and control measures are important tasks for ensuring the landing safety of an aircraft. In recent years, some airports have improved risk prevention and control capabilities to a certain extent through detection technology and intelligent driving technology application. Specifically, in the prior art, the radar detection equipment is mainly used for detecting and tracking in a large range, the optical equipment is guided to identify, and finally bird repelling measures are implemented on birds which are close to the take-off and landing channel of the aircraft. The prior art finds and deals with the basic principle of high risk bird activity correct, but there are some short plates in implementation details, as follows: (1) In the aspect of bird activity monitoring, the ultra-low altitude radar detection tracking performance below 100 meters is obviously reduced, and a large number of birds are active in the space, so that a supplementary detection means is needed. (2) In terms of high risk bird activity judgment, the bird size and runway distance are not accurate enough as the sole criteria for assessing risk. Different bird species, or birds of the same species, have significant differences in the risk of bird strike evolution process in different modes of activity. Therefore, the risk judgment is fine, the bird species and the activity mode are required to be finely identified, and more effective risk pre-judgment and prevention and control methods are provided in a targeted manner. Disclosure of Invention Aiming at the defects of the prior art in the background technology, the embodiment of the invention aims to provide a method, a control center and a system for fine classification and risk prevention and control of fusion detection of airport birds. To achieve the above object, in a first aspect, an embodiment of the present invention provides a method for fine classification and risk prevention and control of fusion detection of birds at an airport, including: The method comprises the steps of acquiring ultra-low altitude bird activity monitoring data through a collaborative detection system, wherein the collaborative detection system comprises radar detection equipment, optical detection equipment and tracking shooting equipment; carrying out fusion processing on the ultralow-altitude bird activity monitoring data, and classifying according to bird species and activity modes to obtain a fine classification result; and carrying out risk prediction and prevention and control according to the fine classification result. As a specific implementation mode of the application, the ultra-low altitude bird activity monitoring data is acquired through the cooperative detection system, and the method specifically comprises the following steps: The radar detection equipment is adopted to perform large-range and ultra-low-altitude bird activity detection to obtain radar detection data; Detecting an ultra-low altitude bird target by adopting optical detection equipment to obtain optical detection data; And guiding the tracking shooting equipment to search for and lock a suspected tracking target according to the radar detection data and the optical detection data to obtain tracking shooting data. As a specific implementation mode of the application, the fine classification result is obtained specifically as follows: The ultra-low altitude bird activity monitoring data is processed by adopting a fusion processing network to obtain classification characteristics, wherein the classification characteristics comprise fusion bird characteristics, micro Doppler characteristics and image characteristics; inputting the classification features into a hierarchical classifier to obtain a fusion classification result; and (5) calling a professional knowledge base to correct the fusion classification result to obtain a fine classification result. As a specific implementation mode of the application, the obtained classification characteristics are specifically as follows: extracting track features from the radar detection data and the optical detection data by adopting the track processing module, inputting the track features into a track classifier, and calculating a track type; Extracting first features from the radar detection data and extracting second features from the radar detection data and the tracking shooting data by adopting the bird feature processing module, wherein the first features