Bird detection and early warning system based on bird detection radar

Abstract

The present invention relates to the technical field of bird detection and early warning, and specifically to a bird detection and early warning system based on a bird detection radar, comprising a data acquisition module, a data analysis module and an early warning module, wherein the data acquisition module acquires bird detection radar data and a dangerous area location; the data analysis module selects an importance distribution according to the bird detection radar data, generates an importance sample according to the importance distribution, and calculates its sample weight; selects samples whose importance samples fall in the dangerous area as dangerous samples, and performs weighted processing according to the corresponding sample weights to obtain the probability that a bird falls in the dangerous area; the early warning module sets a baseline probability threshold, and if the probability that a bird falls in the dangerous area is less than the baseline probability threshold, calculates the expected loss according to the dangerous sample and the corresponding sample weight; determines the loss fluctuation of each dangerous sample according to the relationship between the dangerous sample and the expected loss, and outputs it in a graphical manner.

Description

Technical Field

[0001] The present invention relates to the technical field of flying bird detection and early warning, and in particular to a flying bird detection and early warning system based on a bird detecting radar. Background Art

[0002] Traditional bird trajectory prediction systems have many technical limitations in terms of prediction accuracy and risk identification.

[0003] First, these systems usually rely on average or frequency calculation methods, ignoring the importance differences between samples, which may lead to inaccurate reflection of the actual threat situation of bird activities. Because the weight of each sample is not taken into account, the system is susceptible to data noise, resulting in biased prediction results. This method cannot effectively identify specific samples that have a greater impact on risk assessment, and may underestimate or overestimate the risk of a specific area.

[0004] In addition, traditional systems often judge risks based on static thresholds during the early warning process, and are unable to flexibly respond to changes in bird activity patterns and environmental conditions, lacking sufficient adaptability. This static threshold is prone to false positives or false negatives, weakening the reliability of the system in different environments. When assessing risks, failure to consider the diversity and dynamics of actual losses is likely to lead to neglect of subtle but critical risk changes in decision-making, and failure to provide management with a comprehensive and reliable basis for decision-making.

[0005] Finally, the traditional system may be rigid in the presentation of output results, making it difficult to intuitively display the relationship between various risk factors, and it is difficult for users to effectively identify and understand the deep-level risk characteristics contained therein. These technical problems combined have led to the lack of effectiveness and data support of the traditional bird trajectory prediction system in complex environments, limiting its ability to provide accurate assessment and response to bird disturbances. Summary of the invention

[0006] The purpose of the present invention is to provide a flying bird detection and early warning system based on a bird detection radar to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a bird detection and early warning system based on a bird detection radar, comprising a data acquisition module, a data analysis module and an early warning module, wherein:

[0008] The data acquisition module acquires bird detection radar data and the location of the dangerous area;

[0009] The data analysis module selects the importance distribution according to the bird detection radar data, generates importance samples according to the importance distribution, and calculates the sample weight; selects the samples whose importance samples fall in the danger zone as danger samples, and performs weighted processing according to the corresponding sample weight to obtain the probability of the flying bird falling in the danger zone;

[0010] The early warning module sets a baseline probability threshold. If the probability of a bird landing in a dangerous area is less than the baseline probability threshold, the expected loss is calculated based on the dangerous sample and the corresponding sample weight. Based on the relationship between the dangerous sample and the expected loss, the loss fluctuation of each dangerous sample is determined, and the loss fluctuation and sample weight corresponding to each dangerous sample are output in a graphical manner.

[0011] As a further improvement of the technical solution, the bird detection radar data in the data acquisition module includes time points and geographic coordinates.

[0012] As a further improvement of the technical solution, the data analysis module includes an importance analysis unit, and the specific process of the importance analysis unit generating importance samples and calculating the sample weights thereof includes:

[0013] Statistical analysis is performed on bird radar data to determine its spatial and temporal distribution characteristics. Based on the time points in the bird radar data, the frequency of flying birds in each time period in the bird radar data is calculated. Based on the geographic coordinates in the bird radar data, a spatial heat map is generated to identify high-density aggregation areas.

[0014] According to the frequency of bird appearance and high-density gathering areas, a Gaussian mixture model is used to fit the probability distribution of bird activities as the original distribution; the initial Gaussian mixture model parameters are set based on the calculated frequency of bird appearance and high-density gathering areas; the Gaussian mixture model parameters are optimized through iterative adjustment using the expectation maximization algorithm;

[0015] According to the location of the dangerous area, the original distribution is probabilistically thickened to increase the distribution density at the dangerous area and generate an importance distribution. The probability optimization method is used to increase the density of the corresponding area by adding weights to the corresponding area in the Gaussian mixture model.

[0016] Randomly generate samples from the generated importance distribution as importance samples;

[0017] For each importance sample , its sample weight The calculation formula is as follows:

[0018] ,in is the probability density of the importance sample under the importance distribution; is the probability density of the importance sample under the original distribution.

[0019] As a further improvement of the technical solution, the data analysis module includes a dangerous sample analysis unit, and the process of the dangerous sample analysis unit calculating the probability of a flying bird landing in a dangerous area specifically includes:

[0020] Define the danger zone location as a rectangle ; For each important sample, determine whether it falls in the danger zone, using the indicator function To mark, if the important sample is located in the rectangle If it is within, it is recorded as 1, otherwise it is 0; the importance sample with the result is regarded as a dangerous sample;

[0021] Dangerous samples and corresponding sample weights, calculate the weighted average of dangerous samples to estimate the probability of birds landing in dangerous areas, and estimate the probability Calculated by the following formula:

[0022] ,in Represents the number of important samples; this formula indicates that the probability of a bird landing in the danger zone is obtained by multiplying the indicator function of the important samples in the danger zone by their sample weights and then dividing by the number of important samples.

[0023] As a further improvement of the technical solution, the early warning module includes a safety determination unit, and the process of calculating the expected loss by the safety determination unit specifically includes:

[0024] Set a baseline probability threshold to determine whether the risk of birds in the danger zone is acceptable;

[0025] If the probability of a bird landing in a dangerous area is less than the baseline probability threshold, the expected loss is calculated based on the dangerous samples and the corresponding sample weights. , the calculation formula is as follows:

[0026] ,in Represents the loss value of each dangerous sample, obtained through the public database.

[0027] As a further improvement of the technical solution, the process of generating the dynamic reference probability threshold by the security determination unit specifically includes:

[0028] The baseline probability threshold is adjusted according to the relationship between the expected loss and the maximum loss to generate a dynamic baseline probability threshold. The adjustment formula is as follows:

[0029] ,in represents the dynamic baseline probability threshold, represents the baseline probability threshold, Represents the maximum loss, obtained directly from the public database.

[0030] As a further improvement of the present technical solution, the early warning module includes a visualization unit, which determines the loss fluctuation of each dangerous sample based on the relationship between the dangerous sample and the expected loss, and outputs the loss fluctuation corresponding to each dangerous sample and the sample weight in a graphical manner. The loss fluctuation of each dangerous sample is obtained by subtracting the loss value of each dangerous sample from the expected loss, and a scatter plot is used for visualization output, wherein the X-axis in the scatter plot represents the expected loss, the Y-axis represents the loss value of each dangerous sample, the bubble size represents the sample weight, and the bubble color represents the loss fluctuation amplitude.

[0031] Compared with the prior art, the present invention has the following beneficial effects:

[0032] The bird detection and early warning system based on bird-detecting radar selects samples whose importance samples fall in the danger zone as dangerous samples, and performs weighted processing according to the corresponding sample weights, so as to accurately calculate the probability of birds falling in the danger zone. This process greatly improves the accuracy of system predictions; by weighted averaging samples of different importance, the system can significantly reduce the impact of noise and more realistically reflect the actual threat of bird activities; at the same time, this weighted processing makes the probability calculation no longer rely solely on the frequency of sample occurrence, but takes into account the importance of each sample and its representativeness in the sample distribution. This method ensures that when assessing risks, more attention is paid to samples that have a greater impact on the overall risk; on this basis, the expected loss is further calculated based on the dangerous samples and the corresponding sample weights. By weighted calculation of the loss of each dangerous sample and summarizing the information of all important samples, the potential economic losses that may be caused by the threat of birds in a certain area can be accurately judged. This weighted method not only makes the risk assessment process more comprehensive, but also provides strong data support for subsequent management and decision-making;

[0033] In addition, based on the relationship between dangerous samples and expected losses, the loss fluctuation of each dangerous sample is determined and output in a graphical form. On this basis, the baseline probability threshold is adjusted through the relationship between expected losses and maximum losses to generate a dynamic baseline probability threshold, which has a positive impact on improving the flexibility and adaptability of the system. By analyzing the loss fluctuation, it is possible to clearly see which samples have loss values ​​that deviate more from expectations, thereby identifying uncertainties and potential risks. At the same time, this information is intuitively presented to users through visualization methods such as scatter plots, which can more quickly capture important information and identify concentrated areas of risk. As the environment or bird activity patterns change, the baseline probability threshold is dynamically adjusted so that the system can respond to changes in actual conditions in a timely manner and accurately judge the risk level of birds landing in dangerous areas. The integration of these functions not only increases the system's adaptability when facing different operating conditions, but also enhances its comprehensiveness and accuracy in prediction, decision-making and early warning, providing users with a more efficient risk management solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the overall module of the present invention;

[0035] Figure 2 It is a schematic diagram of a data analysis module of the present invention;

[0036] Figure 3 It is a schematic diagram of the early warning module of the present invention.

[0037] In the figure: 100, data acquisition module; 200, data analysis module; 201, importance analysis unit; 202, dangerous sample analysis unit; 300, early warning module; 301, safety judgment unit; 302, visualization unit. DETAILED DESCRIPTION

[0038] 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.

[0039] Next, see Figure 1-Figure 3 The present invention provides a technical solution: a flying bird detection and early warning system based on a bird detection radar, comprising a data acquisition module 100, a data analysis module 200 and an early warning module 300.

[0040] The data acquisition module 100 acquires bird detection radar data and the location of the dangerous area, wherein the bird detection radar data includes but is not limited to time points and geographic coordinates; the location of the dangerous area is defined by a static configuration file, including fixed area coordinates (polygonal area, vertex coordinates of the polygon);

[0041] The importance analysis unit 201 in the data analysis module 200 selects the importance distribution according to the bird detection radar data, generates importance samples according to the importance distribution, and calculates the sample weights thereof, which specifically includes:

[0042] Statistical analysis is performed on bird radar data to determine its spatial and temporal distribution characteristics. Based on the time points in the bird radar data, the frequency of bird appearance in each time period (e.g., every hour) in the bird radar data is calculated. Based on the geographic coordinates in the bird radar data, a spatial heat map is generated to identify high-density aggregation areas.

[0043] According to the frequency of bird appearance and high-density aggregation areas, a distribution model (such as a Gaussian mixture model) is used to fit the probability distribution of bird activities as the original distribution; the Gaussian mixture model is used because it can effectively capture multi-peak characteristics and complex distribution patterns. Based on the calculated frequency of bird appearance and high-density aggregation areas, initial parameters such as the mean, variance, and weight of each component of the distribution are set; the expectation maximization algorithm is used to iteratively adjust and optimize the model parameters so that the model has a high degree of fit and accurately characterizes the probability density of bird distribution in different times and regions; the final output of this process is an original probability distribution that fully reflects the characteristics of the data, which can capture the multi-period and multi-regional characteristics of bird activities and provide a basis for the generation of subsequent importance distribution;

[0044] According to the location of the dangerous area, the original distribution is probability thickened to increase the distribution density at the dangerous area and generate the importance distribution. In order to thicken the probability of the dangerous area, local adjustments are made on the basis of the original distribution. Specifically, the probability optimization method is used to increase the density of the corresponding areas in the Gaussian mixture model by adding weights to the corresponding areas, so that the probability of these areas in the model is denser.

[0045] Randomly generate samples from the generated importance distribution (i.e., the adjusted probability distribution) as importance samples. Since the importance distribution is adjusted through probability, the generated importance samples will be more frequently concentrated in the dangerous area and high-frequency activity period;

[0046] For each importance sample , its sample weight The calculation formula is as follows:

[0047] ,in is the probability density of the importance sample under the importance distribution; is the probability density of the importance sample under the original distribution.

[0048] The dangerous sample analysis unit 202 in the data analysis module 200 selects samples whose importance samples fall in the dangerous area as dangerous samples, and performs weighted processing according to the corresponding sample weights to obtain the probability of the bird falling in the dangerous area, specifically including:

[0049] Define the danger zone location as a rectangle , for example, the range is from (4,4) to (6,6); for each importance sample, determine whether it falls within the danger zone, using the indicator function To mark, if the important sample is located in the rectangle If the result is within , it is recorded as 1, otherwise it is recorded as 0; the importance sample with the result of 1 is regarded as a dangerous sample;

[0050] Dangerous samples and corresponding sample weights, calculate the weighted average of dangerous samples to estimate the probability of birds landing in dangerous areas, and estimate the probability Calculated by the following formula:

[0051] ,in Represents the number of important samples; this formula indicates that the probability of a bird landing in the danger zone is obtained by multiplying the indicator function of the important samples in the danger zone by their sample weights and then dividing by the number of important samples.

[0052] The safety determination unit 301 in the early warning module 300 sets a reference probability threshold. If the probability of a bird landing in a dangerous area is less than the reference probability threshold, the expected loss is calculated based on the dangerous samples and the corresponding sample weights, specifically including:

[0053] Set a baseline probability threshold to determine whether the risk of birds entering the danger zone is acceptable, for example, 0.10 (i.e., a probability of more than 10% entering the danger zone is considered risky);

[0054] If the probability of a bird landing in a dangerous area is less than the baseline probability threshold, the expected loss is calculated based on the dangerous samples and the corresponding sample weights. , the calculation formula is as follows:

[0055] ,in Represent the loss value for each hazard sample, obtained from historical economic losses, repair costs or other related expenses in public databases. The losses may come from insurance claim records, financial reports or detailed records directly measuring the consequences of accidents;

[0056] In order to improve the flexibility and accuracy of risk assessment and to cope with potential high-risk events in different scenarios, the baseline probability threshold is adjusted according to the relationship between expected loss and maximum loss to generate a dynamic baseline probability threshold. The adjustment formula is as follows:

[0057] ,in represents the dynamic baseline probability threshold, represents the baseline probability threshold, Represents the maximum loss, obtained directly from the public database.

[0058] The visualization unit 302 in the early warning module 300 determines the loss fluctuation of each dangerous sample according to the relationship between the dangerous sample and the expected loss, and outputs the loss fluctuation corresponding to each dangerous sample and the sample weight in a graphical manner. The loss fluctuation of each dangerous sample is obtained by subtracting the loss value of each dangerous sample from the expected loss. In order to draw the relationship between the loss value of each dangerous sample and the expected loss value, a scatter plot is used for visualization output, wherein the X-axis in the scatter plot represents the expected loss, the Y-axis represents the loss value of each dangerous sample, the bubble size represents the sample weight, and the bubble color represents the loss fluctuation amplitude (using a color mapping table, red represents high fluctuation, and blue represents low fluctuation).

[0059] 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. A bird detection and early warning system based on bird detection radar, characterized in that: It comprises a data acquisition module (100), a data analysis module (200) and an early warning module (300), wherein: The data acquisition module (100) acquires bird detection radar data and the location of the dangerous area; The data analysis module (200) selects importance distribution according to the bird detection radar data, generates importance samples according to the importance distribution, and calculates the sample weight; selects samples whose importance samples fall in the danger zone as danger samples, and performs weighted processing according to the corresponding sample weight to obtain the probability of the flying bird falling in the danger zone; The early warning module (300) sets a reference probability threshold, and if the probability of a bird landing in a dangerous area is less than the reference probability threshold, the expected loss is calculated based on the dangerous sample and the corresponding sample weight; based on the relationship between the dangerous sample and the expected loss, the loss fluctuation of each dangerous sample is determined, and the loss fluctuation and sample weight corresponding to each dangerous sample are output in a graphical manner; wherein: The early warning module (300) comprises a safety determination unit (301), and the process of calculating the expected loss by the safety determination unit (301) specifically comprises: Set a baseline probability threshold to determine whether the risk of birds in the danger zone is acceptable; If the probability of a bird landing in a dangerous area is less than the baseline probability threshold, the expected loss is calculated based on the dangerous samples and the corresponding sample weights. , the calculation formula is as follows: ,in It represents the loss value of each dangerous sample, obtained through the public database; The process of generating the dynamic reference probability threshold by the safety determination unit (301) specifically includes: The baseline probability threshold is adjusted according to the relationship between the expected loss and the maximum loss to generate a dynamic baseline probability threshold. The adjustment formula is as follows: ,in represents the dynamic baseline probability threshold, represents the baseline probability threshold, Represents the maximum loss, obtained directly from the public database.

2. The bird detection and early warning system based on bird detection radar according to claim 1 is characterized in that: The bird detection radar data in the data acquisition module (100) includes a time point and a geographical coordinate.

3. The bird detection and early warning system based on bird detection radar according to claim 2 is characterized in that: The data analysis module (200) comprises an importance analysis unit (202), and the specific process of the importance analysis unit (202) generating importance samples and calculating the sample weights thereof comprises: Statistical analysis is performed on bird radar data to determine its spatial and temporal distribution characteristics. Based on the time points in the bird radar data, the frequency of flying birds in each time period in the bird radar data is calculated. Based on the geographic coordinates in the bird radar data, a spatial heat map is generated to identify high-density aggregation areas. According to the frequency of bird appearance and high-density gathering areas, a Gaussian mixture model is used to fit the probability distribution of bird activities as the original distribution; the initial Gaussian mixture model parameters are set based on the calculated frequency of bird appearance and high-density gathering areas; the Gaussian mixture model parameters are optimized through iterative adjustment using the expectation maximization algorithm; According to the location of the dangerous area, the original distribution is probabilistically thickened to increase the distribution density at the dangerous area and generate an importance distribution. The probability optimization method is used to increase the density of the corresponding area by adding weights to the corresponding area in the Gaussian mixture model. Randomly generate samples from the generated importance distribution as importance samples; For each importance sample , its sample weight The calculation formula is as follows: ,in is the probability density of the importance sample under the importance distribution; is the probability density of the importance sample under the original distribution.

4. The bird detection and early warning system based on bird detection radar according to claim 3 is characterized in that: The data analysis module (200) comprises a dangerous sample analysis unit (201), and the process of the dangerous sample analysis unit (201) calculating the probability of a flying bird landing in a dangerous area specifically comprises: Define the danger zone location as a rectangle ; For each important sample, determine whether it falls in the danger zone, using the indicator function To mark, if the important sample is located in the rectangle If the result is within , it is recorded as 1, otherwise it is recorded as 0; the importance sample with the result of 1 is regarded as a dangerous sample; Dangerous samples and corresponding sample weights, calculate the weighted average of dangerous samples to estimate the probability of birds landing in dangerous areas, and estimate the probability Calculated by the following formula: ,in Represents the number of important samples; this formula indicates that the probability of a bird landing in the danger zone is obtained by multiplying the indicator function of the important samples in the danger zone by their sample weights and then dividing by the number of important samples.

5. The bird detection and early warning system based on bird detection radar according to claim 4 is characterized in that: The early warning module (300) includes a visualization unit (302), which determines the loss fluctuation of each dangerous sample based on the relationship between the dangerous sample and the expected loss, outputs the loss fluctuation corresponding to each dangerous sample and the sample weight in a graphical manner, obtains the loss fluctuation of each dangerous sample by subtracting the loss value of each dangerous sample from the expected loss, and uses a scatter plot for visualization output, wherein the X-axis in the scatter plot represents the expected loss, the Y-axis represents the loss value of each dangerous sample, the bubble size represents the sample weight, and the bubble color represents the loss fluctuation amplitude.

Images