A bird detection method, system and device for airport based on bird detection radar

Abstract

The present invention relates to the field of bird detection technology, and specifically to an airport bird detection method, system and equipment based on bird detection radar. In the present invention, a data collection module collects data from a bird detection radar, including historical and real-time bird data, wherein the historical bird data includes bird types, bird behavior patterns and historical bird characteristics; a model construction module uses a decision tree algorithm to split the historical bird characteristics according to the historical bird data to form a primary decision tree, in which leaf nodes identify the bird types, and then uses these leaf nodes as root nodes to construct a secondary decision tree, which is further refined into a bird detection model; this two-layer decision tree structure effectively captures the diversity of bird species and behaviors; a model application module uses a bird detection model to predict the bird types and behavior patterns in the real-time bird data, and generates a risk level by comparing the characteristic values ​​of the same path nodes with a safety threshold.

Description

Technical Field

[0001] The present invention relates to the technical field of bird detection, and in particular to an airport bird detection method, system and equipment based on bird detection radar. Background Art

[0002] The warning of traditional bird detection radar mainly relies on the prediction of bird trajectories, but this method is often insufficient when identifying and responding to complex bird behaviors. Because different types of birds have their own unique flight patterns and behavioral characteristics, relying solely on trajectory information to determine whether birds will interfere with airport operations is prone to misjudgment and omission. Inaccurate trajectory prediction, especially when bird behavior is changeable, will lead to a lag in safety measures. Therefore, a comprehensive analysis model that can combine bird species and behavior patterns is needed to improve the accuracy of trajectory prediction and provide reliable technical support for airport operation safety. Through this multi-dimensional analysis, the system can better understand bird behavior and make quick and effective decisions.

[0003] On the other hand, in traditional airport bird detection, real-time bird data usually contains a large amount of feature information, and effective prediction and response often rely on the correct selection of these data features. Due to the lack of a systematic feature selection mechanism, traditional methods are easily troubled by information overload when processing real-time data, and key features are easily submerged in too much irrelevant information, resulting in reduced detection efficiency and accuracy. To address this problem, existing methods can use the path nodes in the decision tree algorithm to perform effective feature selection. Through this strategy, the most influential key variables can be extracted from the numerous data features, thereby conducting more accurate analysis and risk assessment. Summary of the invention

[0004] The purpose of the present invention is to provide an airport bird detection method, system and equipment based on bird detection radar to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, one of the objectives of the present invention is to provide an airport bird detection system based on bird detection radar, comprising a data collection module, a model building module and a model application module, wherein:

[0006] The data collection module collects bird detection radar data, which includes historical bird flying data and real-time bird flying data; wherein the historical bird flying data includes bird types, bird behavior patterns and historical bird flying characteristics;

[0007] The model building module uses a decision tree algorithm to split the path nodes of the primary decision tree for the historical bird features according to the bird types in the historical bird data, wherein the path nodes include leaf nodes, and the leaf nodes represent the bird types; the leaf nodes are used as the root nodes of the secondary decision tree, and the path nodes of the secondary decision tree are split for the historical bird features according to the bird behavior patterns in the historical bird data, and a bird detection model is formed;

[0008] The model application module predicts the bird type and behavior pattern of the real-time bird data based on the constructed bird detection model; and selects the same path nodes in the primary decision tree and the secondary decision tree in the prediction process, compares the numerical value corresponding to the data feature in the path node with the safety threshold, and determines the risk level of the bird under the current behavior model.

[0009] As a further improvement of the present technical solution, the model construction module includes a primary decision tree construction unit, and the process of the primary decision tree construction unit constructing a primary decision tree specifically includes:

[0010] Extract historical bird data Includes a subset of historical bird features and bird types, indicating ;

[0011] For historical bird features For each feature in , calculate the information gain, and the calculation formula is as follows:

[0012] ,in Representation dataset Using the History Flyer feature Information gain after splitting; It is a dataset The entropy of It is a dataset Entropy, which is a measure of the uncertainty of a data set; Indicates historical bird characteristics The set of all values ​​of ; Indicates the eigenvalue When , the number of samples in the corresponding data set; express The number of samples in the dataset;

[0013] The historical bird features with the largest information gain are selected for splitting to form path nodes, and the tree structure is recursively constructed until each leaf node determines a unique bird type.

[0014] As a further improvement of the present technical solution, the model construction module includes a secondary decision tree construction unit, and the process of the secondary decision tree construction unit constructing the secondary decision tree specifically includes:

[0015] Extract historical bird data Includes a subset of historical bird features and bird behavior patterns and the leaf nodes of each primary decision tree The type of bird represented by express;

[0016] Similar to the primary decision tree, for each historical bird feature The information gain is calculated again to predict the behavior pattern. The calculation formula is as follows:

[0017] ,in Representation dataset Using the History Flyer feature Information gain after splitting;

[0018] The historical bird features with the largest information gain are selected for splitting to form the path nodes of the secondary decision tree, until each leaf node in the secondary decision tree represents a bird behavior pattern, forming a behavior pattern classification leaf node, and forming a bird detection model.

[0019] As a further improvement of the technical solution, the process of determining the risk level by the model application module specifically includes:

[0020] Real-time bird data Input the data into the constructed primary decision tree, predict the type of bird in real-time, and record the primary decision tree path nodes passed;

[0021] Based on the predicted bird type, select the corresponding secondary decision tree, predict the bird's behavior pattern, and record the secondary decision tree path nodes passed;

[0022] Select the same path nodes of the primary decision tree and the secondary decision tree in the prediction process as the key path nodes, and extract the data features and corresponding feature values ​​of each node from the key path nodes;

[0023] The safety threshold of each data feature in the historical bird features is obtained from the public rule base, the feature value in the key path node is compared with the corresponding safety threshold, and the risk level of each data feature in the key path node is generated.

[0024] The second object of the present invention is to provide a method for using an airport bird detection system based on a bird detection radar, comprising the following method steps:

[0025] S1. Collect bird radar data, which includes historical bird data and real-time bird data; the historical bird data includes bird types, bird behavior patterns and historical bird characteristics;

[0026] S2. According to the types of birds in the historical bird data, the decision tree algorithm is used to split the path nodes of the primary decision tree for the historical bird features, wherein the path nodes include leaf nodes, and the leaf nodes represent the types of birds; the leaf nodes are used as the root nodes of the secondary decision tree, and according to the bird behavior patterns in the historical bird data, the path nodes of the secondary decision tree are split for the historical bird features, and a bird detection model is formed;

[0027] S3. According to the constructed bird detection model, the bird type and bird behavior pattern of the real-time bird data are predicted; and the same path nodes of the primary decision tree and the secondary decision tree are selected in the prediction process, and the numerical value corresponding to the data feature in the path node is compared with the safety threshold to determine the risk level of the bird under the current behavior model.

[0028] The third object of the present invention is to provide an airport bird detection device based on a bird-detecting radar, comprising: a memory and a processor; the memory is used to store program instructions; the processor is used to call and execute the program instructions in the memory.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The airport bird detection method, system and equipment based on bird detection radar realize multiple applications of path nodes by innovatively combining primary and secondary decision trees, and significantly improve the prediction accuracy of bird types and behavior patterns. First, the system uses the bird types in historical bird data to construct a primary decision tree. The path node splits based on the historical bird features that maximize information gain, thereby ensuring that each leaf node represents a unique bird type. This hierarchical tree structure method not only effectively captures the diversity of bird species, but also provides a good classification basis. Then, the system uses the leaf nodes of the primary decision tree as the root nodes of the secondary decision tree, and further splits the path nodes on the historical bird features through the analysis of the bird behavior patterns to form a complete bird detection model. This enables the system to use the model to accurately predict the bird type and its behavior pattern when processing real-time bird data, and compare the data features of the same path nodes in the primary and secondary decision trees with the safety threshold, thereby determining the risk level of the bird under the current behavior model. This refined path node application provides a systematic approach from data characteristics to risk levels, providing precise and personalized support for airport bird monitoring.

[0031] At the same time, this method effectively solves the limitations of traditional bird detection radar warnings, namely, relying solely on trajectories to predict whether birds will interfere with airports, which often leads to misjudgments due to differences in species and behavior patterns; by constructing and applying decision trees that are subdivided into path nodes, the system no longer relies solely on trajectories, but uses multi-level data analysis, including two important variables: bird species and behavior patterns, to make predictions more accurate; in addition, in the face of many data features in real-time bird data, the system selects important data features through a decision tree algorithm, eliminates the interference of a large amount of irrelevant data, and focuses on the data features of key path nodes; this feature selection mechanism provides clear indicators for subsequent risk judgments, and produces reliable risk assessment results by comparing with safety thresholds, thereby improving the system's response speed and decision-making effectiveness, and providing a more scientific basis for judgment for airport safety operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0033] Figure 2 It is a schematic diagram of the overall module unit of the present invention.

[0034] In the figure: 100, data collection module; 200, model construction module; 201, primary decision tree construction unit; 202, secondary decision tree construction unit; 300, model application module. DETAILED DESCRIPTION

[0035] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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.

[0036] Next, see Figure 1-Figure 2 One of the purposes of this embodiment is to provide an airport bird detection system based on bird detection radar, including a data collection module 100, a model building module 200 and a model application module 300.

[0037] The data collection module 100 collects bird radar data, which includes historical bird data and real-time bird data; the historical bird data includes bird types, bird behavior patterns (such as foraging, migration, nesting, etc.) and historical bird characteristics (such as size, flight speed and direction, etc.); while the real-time bird data does not include known bird types and bird behavior patterns. Indicates that Indicates the historical bird characteristics, Indicates the type of bird. Indicates bird behavior patterns; real-time bird data is used Indicates that Indicates real-time bird-flying features.

[0038] The primary decision tree construction unit 201 in the model construction module 200 uses a decision tree algorithm to split the primary decision tree path nodes of the historical bird features according to the bird types in the historical bird data, wherein the path nodes include leaf nodes, and the leaf nodes represent the bird types, specifically including:

[0039] Extract historical bird data Includes a subset of historical bird features and bird types, indicating ;

[0040] For historical bird features For each feature (such as size, flight speed, and direction), calculate the information gain using the following formula:

[0041] ,in Representation dataset Using the History Flyer feature The information gain after splitting indicates the degree of uncertainty reduction after the splitting due to the introduction of a historical bird feature; It is a dataset The entropy of It is a dataset Entropy, entropy is an indicator to measure the uncertainty of the data set. The higher the entropy, the greater the uncertainty of the classification; Indicates historical bird characteristics The set of all values ​​of ; Indicates the eigenvalue When , the number of samples in the corresponding data set; express The number of samples in the dataset;

[0042] The historical bird features with the largest information gain are selected for splitting to form path nodes, and the tree structure is recursively constructed until each leaf node determines a unique bird type or until the set stopping condition (such as the maximum depth of the tree) is reached.

[0043] The secondary decision tree construction unit 202 in the model construction module 200 uses the leaf node as the root node of the secondary decision tree, splits the secondary decision tree path node for the historical bird features according to the bird behavior pattern in the historical bird data, and forms a bird detection model, which specifically includes:

[0044] Extract historical bird data Includes a subset of historical bird features and bird behavior patterns and the leaf nodes of each primary decision tree The type of bird represented by express;

[0045] Similar to the primary decision tree, for each historical bird feature The information gain is calculated again to predict the behavior pattern. The calculation formula is as follows:

[0046] ,in Representation dataset Using the History Flyer feature Information gain after splitting;

[0047] The historical bird features with the largest information gain are selected for splitting to form the path nodes of the secondary decision tree, until each leaf node in the secondary decision tree represents a bird behavior pattern, forming a behavior pattern classification leaf node, and forming a bird detection model.

[0048] The model application module 300 predicts the type of bird and the behavior pattern of the bird in real-time bird data according to the constructed bird detection model; and selects the same path node in the primary decision tree and the secondary decision tree in the prediction process, and compares the value corresponding to the data feature in the path node with the safety threshold to determine the risk level of the bird under the predicted bird behavior pattern, specifically including:

[0049] Real-time bird data Input the data into the constructed primary decision tree, predict the type of bird in real-time, and record the primary decision tree path nodes passed;

[0050] Based on the predicted bird type, select the corresponding secondary decision tree, predict the bird's behavior pattern, and record the secondary decision tree path nodes passed;

[0051] Select the same path nodes of the primary decision tree and the secondary decision tree in the prediction process as the key path nodes, and extract the data features and corresponding feature values ​​of each node from the key path nodes;

[0052] The safety threshold of each data feature in the historical bird features is obtained from the public rule base, the feature value in the critical path node is compared with the corresponding safety threshold, and the risk level of each data feature in the critical path node is generated; the management personnel will implement the corresponding response mechanism according to the risk level.

[0053] The second purpose of this embodiment is to provide a method for using a flying bird distance assessment system based on a bird detection radar, comprising the following method steps:

[0054] S1. Collect bird radar data, which includes historical bird data and real-time bird data; the historical bird data includes bird types, bird behavior patterns and historical bird characteristics;

[0055] S2. According to the types of birds in the historical bird data, the decision tree algorithm is used to split the path nodes of the primary decision tree for the historical bird features, wherein the path nodes include leaf nodes, and the leaf nodes represent the types of birds; the leaf nodes are used as the root nodes of the secondary decision tree, and according to the bird behavior patterns in the historical bird data, the path nodes of the secondary decision tree are split for the historical bird features, and a bird detection model is formed;

[0056] S3. According to the constructed bird detection model, the bird type and bird behavior pattern of the real-time bird data are predicted; and the same path nodes of the primary decision tree and the secondary decision tree are selected in the prediction process, and the numerical value corresponding to the data feature in the path node is compared with the safety threshold to determine the risk level of the bird under the current behavior model.

[0057] The third purpose of this embodiment is to provide an airport bird detection device based on a bird-detecting radar, comprising: a memory and a processor; the memory is used to store program instructions; the processor is used to call and execute the program instructions in the memory.

[0058] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An airport bird detection system based on bird detection radar, characterized in that: The method comprises a data collection module (100), a model building module (200) and a model application module (300), wherein: The data collection module (100) collects bird detection radar data, the bird detection radar data including historical bird flying data and real-time bird flying data; wherein the historical bird flying data includes bird types, bird behavior patterns and historical bird flying characteristics; The model building module (200) uses a decision tree algorithm to split the path nodes of a primary decision tree for historical bird features according to the bird types in the historical bird data, wherein the path nodes include leaf nodes, and the leaf nodes represent the bird types; uses the leaf nodes as the root nodes of a secondary decision tree, and splits the path nodes of the secondary decision tree for historical bird features according to the bird behavior patterns in the historical bird data, and forms a bird detection model; wherein: The model construction module (200) comprises a primary decision tree construction unit (201), and the process of constructing a primary decision tree by the primary decision tree construction unit (201) specifically comprises: Extract historical bird data Includes a subset of historical bird features and bird types, indicating ; For historical bird features For each feature in , calculate the information gain, and the calculation formula is as follows: ,in Representation dataset Using the History Flyer feature Information gain after splitting; It is a dataset The entropy of It is a dataset Entropy, which is a measure of the uncertainty of a data set; Indicates historical bird characteristics The set of all values ​​of ; Indicates the eigenvalue When , the number of samples in the corresponding data set; express The number of samples in the dataset; Select the historical bird features with the largest information gain for splitting, forming path nodes, and recursively build the tree structure until each leaf node determines a unique bird type; The model construction module (200) comprises a secondary decision tree construction unit (202), and the process of constructing a secondary decision tree by the secondary decision tree construction unit (202) specifically comprises: Extract historical bird data Includes a subset of historical bird features and bird behavior patterns and the leaf nodes of each primary decision tree The type of bird represented by express; Similar to the primary decision tree, for each historical bird feature The information gain is calculated again to predict the behavior pattern. The calculation formula is as follows: ,in Representation dataset Using the History Flyer feature Information gain after splitting; The historical bird features with the largest information gain are selected for splitting to form the path nodes of the secondary decision tree, until each leaf node in the secondary decision tree represents a bird behavior pattern, forming a behavior pattern classification leaf node, and forming a bird detection model; The model application module (300) predicts the type of bird and the behavior pattern of the bird based on the constructed bird detection model. It selects the same path node in the primary decision tree and the secondary decision tree in the prediction process, compares the value corresponding to the data feature in the path node with the safety threshold, and determines the risk level of the bird under the current behavior model.

2. The airport bird detection system based on bird detection radar according to claim 1 is characterized in that: The process of determining the risk level by the model application module (300) specifically includes: Real-time bird data Input the data into the constructed primary decision tree, predict the type of bird in real-time, and record the primary decision tree path nodes passed; Based on the predicted bird type, select the corresponding secondary decision tree, predict the bird's behavior pattern, and record the secondary decision tree path nodes passed; Select the same path nodes of the primary decision tree and the secondary decision tree in the prediction process as the key path nodes, and extract the data features and corresponding feature values ​​of each node from the key path nodes; The safety threshold of each data feature in the historical bird features is obtained from the public rule base, the feature value in the key path node is compared with the corresponding safety threshold, and the risk level of each data feature in the key path node is generated.

3. A method for using an airport bird detection system based on a bird detection radar according to any one of claims 1 to 2, characterized in that: The method comprises the following steps: S1. Collect bird radar data, which includes historical bird data and real-time bird data; the historical bird data includes bird types, bird behavior patterns and historical bird characteristics; S2. According to the types of birds in the historical bird data, the decision tree algorithm is used to split the path nodes of the primary decision tree for the historical bird features, wherein the path nodes include leaf nodes, and the leaf nodes represent the types of birds; the leaf nodes are used as the root nodes of the secondary decision tree, and according to the bird behavior patterns in the historical bird data, the path nodes of the secondary decision tree are split for the historical bird features, and a bird detection model is formed; S3. Predict the bird type and bird behavior pattern of the real-time bird data based on the constructed bird detection model; The same path nodes of the primary decision tree and the secondary decision tree in the prediction process are selected, and the numerical values ​​corresponding to the data features in the path nodes are compared with the safety threshold to determine the risk level of the bird under the current behavior model.

4. An airport bird detection device based on bird detection radar, characterized in that: include: Memory and processor; The memory is used to store program instructions; The processor is used to call and execute the program instructions in the memory to perform the method according to claim 3.

Images