An airborne air pollution detection method based on harmonic vector variation

Abstract

The application discloses an airborne air pollution detection method based on harmonic vector change and belongs to the technical field of unmanned aerial vehicle (UAV) pollution detection. The steps are as follows: a harmonic initialization model is constructed, and the cosine, sine function amplitude and the pollution background value baseline of a to-be-detected area are solved to obtain a harmonic model of the to-be-detected area; the measurement value at the current time is obtained through a sensor module; the current time is substituted into the harmonic model of the to-be-detected area to obtain the prediction value at the current time; the abnormality judgment index is obtained based on the ratio method and compared with an abnormality threshold value; when N abnormal time instants are continuously detected, the vector change angle pollution judgment is performed to distinguish the UAV noise influence condition and the air pollution condition. The method provided by the application can more sensitively and accurately judge the pollution condition of the to-be-detected area.

Description

Technical Field

[0001] The present invention belongs to the technical field of pollution detection by unmanned aerial vehicles (UAVs), and in particular relates to an airborne air pollution detection method based on harmonic vector changes. Background Art

[0002] With the rapid development of urbanization and industrialization, the number of heavily polluting industrial enterprises has increased rapidly, causing significant pollution to the ambient air. Currently, researchers are working to develop drones for environmental monitoring to enhance the flexibility of air quality monitoring. Implementing drone monitoring and early warning in chemical industry scenarios is a challenging problem, and the solution involves both hardware and software.

[0003] On the hardware side, various high-precision sensors and monitoring devices are utilized. These devices often have built-in intelligent gas processing algorithms to achieve precise conversion of digital signals to analog signals. On the software side, algorithms can be designed to reduce noise interference. Drone flight is often accompanied by high-speed airflow and high-frequency vibrations of the fuselage. The presence of a large amount of noise makes it difficult for onboard sensors to accurately detect and provide early warnings. Specifically, onboard sensors experience significant inaccuracies when measuring gases, and their measured values ​​often fluctuate widely. Furthermore, drones are limited by battery power and cannot fly for long periods of time. How to quickly eliminate the impact of this noise and accurately detect pollution levels in chemical plants, thereby achieving effective early warning, is an urgent problem that needs to be solved. Summary of the Invention

[0004] The purpose of the present invention is to address the deficiencies of the prior art and provide an airborne air pollution detection method based on harmonic vector changes.

[0005] The specific technical solutions adopted in the present invention are as follows:

[0006] The present invention provides an airborne air pollution detection method based on harmonic vector changes. The air pollution detection method utilizes an unmanned aerial vehicle (UAV) device and a host computer to determine the gas pollution status in a test area. The UAV device includes an aircraft module, a sensor module, and a communication module provided on the aircraft module. Information collected by the sensor module is transmitted to the host computer via the communication module. The specific steps for performing air pollution detection in the host computer are as follows:

[0007] S1: Construct a harmonic initialization model; the aircraft is at a specified altitude L, with the sensor measurements obtained as the dependent variable and time as the independent variable, and the cosine function amplitude and sine function amplitude in the harmonic initialization model are fitted and solved;

[0008] S2: Substitute the cosine function amplitude, sine function amplitude and pollution background value baseline of the area to be measured obtained in step S1 into the harmonic initialization model to construct a harmonic model of the area to be measured;

[0009] S3: After the aircraft reaches the target altitude H, the sensor module obtains the current measurement value Y(t); the current time is substituted into the harmonic model of the measured area constructed in step S2 to obtain the predicted value P(t) at the current time;

[0010] S4: Based on the difference between the current measured value Y(t) and the current predicted value P(t) and the root mean square error of the harmonic model at the previous moment, an abnormality judgment indication is obtained; if the abnormality judgment indication is less than the abnormality threshold, the current moment is determined to be a normal moment; if the abnormality judgment indication is greater than the abnormality threshold, the current moment is determined to be an abnormal moment;

[0011] S5: When N abnormal moments are detected continuously, perform vector change angle pollution judgment; calculate the vector change angle of every two adjacent measurement values ​​at the abnormal moment, and then calculate the average value of the vector change angle; if the average value of the vector change angle is greater than 30°, it is judged that the abnormal moment is affected by drone noise; if the average value of the vector change angle is ≤30°, it is judged that the abnormal moment is affected by air pollution.

[0012] Preferably, the harmonic initialization model is as follows:

[0013]

[0014] Where: P(t)′ is the sensor measurement value, b0 is the pollution background value baseline of the measured area; a1 is the amplitude of the cosine function; b1 is the amplitude of the sine function; T is the vibration period of the cosine function or sine function.

[0015] Furthermore, the pollution background value baseline b0 of the test area is measured as follows: the aircraft is located in the test area, and the sensor module is turned on for preheating when not in flight; after the sensor module reaches a stable measurement state, data is collected for T0 minutes and the data obtained by the sensor module is transmitted to the host computer through the communication module; the average of the data obtained by the sensor module within T0 minutes is the pollution background value baseline b0 of the test area;

[0016] The fitting and solving process of the cosine function amplitude a1 and the sine function amplitude b1 in the harmonic initialization model is as follows: the aircraft is located in the test area, flies to a specified altitude L and maintains it, turns on the sensor module to collect data for T1 minutes and transmits the sensor measurement value P(t)′ to the host computer through the communication module; uses the sensor measurement value P(t)′ as the dependent variable and time as the independent variable to fit and solve the cosine function amplitude a1 and the sine function amplitude b1 in the harmonic initialization model.

[0017] Furthermore, the designated height L=3 m; the acquisition time T0 and T1 are both 5 minutes.

[0018] As preferred, the to-be-tested area harmonic model is as follows:

[0019]

[0020] In the formula, P(t) is a prediction value at a current time; t is the current time; the target height H of the aircraft flight is greater than 3 m.

[0021] As preferred, the abnormality judgment index n in step S4 is calculated as follows:

[0022]

[0023] In the formula, delta(t) represents a difference between the measurement value Y(t) at the current time and the prediction value P(t) at the current time, and RMSE(t-1) represents a root mean square error of the harmonic model at the previous time.

[0024] As preferred, the abnormality threshold n0 in step S4 is 4, representing that the absolute value of the difference between the prediction value and the sensor is greater than four times the root mean square error.

[0025] As preferred, when five abnormal time points are continuously detected in step S5, the vector change angle pollution judgment is performed.

[0026] As preferred, the vector change angle pollution judgment in step S5 is specifically as follows:

[0027] In the coordinate system, it is assumed that the value of the first abnormal time point is the origin, the measurement value at the i-th abnormal time point (1 < i < 5) and the measurement value at the first abnormal time point form a vector, the measurement value at the i+1-th abnormal time point and the measurement value at the first abnormal time point form a vector, and the two vectors form an angle β i，i+1 By analogy; the average value of the vector change angle

[0028] As preferred, the sensor module is a carbon dioxide monitoring sensor, a sulfur dioxide monitoring sensor or a nitrogen dioxide monitoring sensor.

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

[0030] The air pollution detection method provided by the present application is based on the gas data collected by the unmanned aerial vehicle in the to-be-tested area, and whether an abnormality occurs is preliminarily determined by the ratio method, and if the abnormality continuously occurs, further pollution judgment is performed based on the vector change angle. Through the vector change angle judgment, the judgment error caused by the unmanned aerial vehicle vibration, air flow, sensor hardware measurement fluctuation and the like can be avoided. The method provided by the present application can more sensitively and accurately judge the pollution condition of the to-be-tested area. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 The schematic diagram of the unmanned aerial vehicle device and the host computer provided in the embodiment is shown in the following figure:

[0032] Figure 2 The flow chart of the airborne air pollution detection method provided in the embodiment is shown in the following figure:

[0033] Figure 3 The schematic diagram of the difference between the measurement value Y(t) at the current time and the predicted value P(t) at the current time in the embodiment is shown in the following figure:

[0034] Figure 4 The schematic diagram of the vector variation calculation of the measurement values at adjacent time points in the embodiment is shown in the following figure:

[0035] Figure 5 The example diagram of the vector variation of the measurement values under different conditions in the embodiment is shown in the following figure. DETAILED DESCRIPTION

[0036] The present application will be further described and explained with reference to the accompanying drawings and specific embodiments. The technical features of each embodiment in the present application can be combined accordingly without conflict.

[0037] As a preferred specific embodiment of the present application, the embodiment provides an airborne air pollution detection method based on harmonic vector variation. The air pollution detection method uses an unmanned aerial vehicle device and a host computer to determine the gas pollution condition of a to-be-detected area. The to-be-detected area in the embodiment is a certain chemical industrial park.

[0038] As shown in the figure, Figure 1 The unmanned aerial vehicle device provided in the embodiment includes an aircraft module, a sensor module arranged on the aircraft module, and a communication module. The aircraft serves as a flight carrier. The sensor module collects and obtains corresponding data of the gas in the air. The communication module transmits the data collected by the sensor module to the host computer through a line.

[0039] In the embodiment, the sensor module is a cuboid box located at the center of the upper surface of the aircraft. The box mainly includes a lithium battery, a circuit, a pump, and a gas sensor. The pump is connected with two gas guide pipes, one gas inlet pipe, and one gas outlet pipe. The sensor module includes a circuit board. All sensors are arranged in a square form at equal intervals to increase heat dissipation and reduce the influence of temperature on the sensors.

[0040] The sensor module includes a gas detection sensor, which in this embodiment includes a carbon dioxide monitoring sensor, a sulfur dioxide monitoring sensor, a nitrogen dioxide monitoring sensor, a temperature monitoring sensor, and a humidity monitoring sensor. The sensor detects the gas and outputs an analog voltage signal. The single-chip microcomputer samples and converts this analog signal through the sampling and conversion module (ADC) to obtain gas concentration or other related data. The single-chip microcomputer transmits the relevant data to the RS485 communication module via a TTL signal. The RS485 communication module is responsible for converting the TTL signal into an RS485 signal and sending the data to the sending station via the RS485 bus. The sending station transmits the data to the receiving station via the RS485 signal. The receiving station receives the signal and uploads it to the host computer via the RS485 to TTL signal conversion.

[0041] After receiving the data, the host computer performs air pollution detection. The process is as follows: Figure 2 The specific steps are as follows:

[0042] (1) Constructing the harmonic initialization model:

[0043]

[0044] Where: P(t)′ is the sensor measurement value, b0 is the pollution background value baseline of the measured area; a1 is the amplitude of the cosine function; b1 is the amplitude of the sine function; T is the vibration period of the cosine function or sine function

[0045] The method for measuring the pollution background value baseline b0 of the area to be measured is as follows: the aircraft is located in the area to be measured, and the sensor module is turned on for preheating when not in flight; after the sensor module reaches a stable measurement state, the data is collected for T0 minutes (5 minutes in this embodiment) and the data obtained by the sensor module is transmitted to the host computer through the communication module, with an acquisition frequency of 1 second / packet. The average value of the data obtained by the sensor module within T0 minutes is the pollution background value baseline b0 of the area to be measured. The b0 of the same area to be measured is a fixed value, and the baseline value usually reflects the basic pollution situation of the area to a certain extent.

[0046] The fitting and solving process of the cosine function amplitude a1 and the sine function amplitude b1 is as follows: the aircraft is located in the test area, flies to a specified altitude L and maintains it, turns on the sensor module to collect data for T1 minutes (5 minutes in this embodiment) and transmits the sensor measurement value P(t)′ to the host computer through the communication module; uses the sensor measurement value P(t)′ as the dependent variable and time as the independent variable to fit and solve the cosine function amplitude a1 and the sine function amplitude b1 in the harmonic initialization model.

[0047] In this embodiment, the designated height L=3 m. The measurement is performed at a height of 3 meters because the air at this height can affect human breathing conditions, and in most cases in chemical parks, the gas conditions measured at this height will not show pollution values.

[0048] (2) Construct a harmonic model of the area to be tested.

[0049] Substitute the cosine function amplitude a1, sine function amplitude b1 and pollution background value baseline b0 of the measured area obtained in step (1) into the harmonic initialization model to obtain the harmonic model of the chemical park:

[0050]

[0051] Where: P(t) is the predicted value at the current moment; t is the current moment.

[0052] (3) Obtain measured and predicted values.

[0053] After the aircraft reaches the target altitude H (H>3m), the sensor module obtains the current measurement value Y(t). The current time is substituted into the harmonic model of the measured area constructed in step (2) to obtain the predicted value P(t) at the current time.

[0054] (4) Make abnormal judgment based on the ratio method.

[0055] The predicted value represents an estimate of the next moment from the previous moment. This value is usually stable. However, when the sensor measurement value changes dramatically at the next moment, this sudden increase could be due to noise or pollutants. In this paper, such a sudden change is referred to as an abnormal moment. To quantify the difference between the predicted value and the sensor measurement value, a ratio method is used for abnormality determination.

[0056] like Figure 3 As shown, the difference delta(t) between the current measured value Y(t) and the current predicted value P(t) is calculated, as well as the root mean square error RMSE(t-1) of the harmonic model at the previous moment, to obtain the abnormal judgment index n:

[0057]

[0058] If the abnormality judgment indication number n is less than the abnormality threshold n0, the current moment is determined to be a normal moment; if the abnormality judgment indication number n is greater than the abnormality threshold n0, the current moment is determined to be an abnormal moment.

[0059] In this embodiment, the abnormal threshold n0 is set to 4, which means that the absolute value of the difference between the predicted value and the sensor value is greater than four times the root mean square error.

[0060] (5) Pollution judgment based on vector change angle.

[0061] When n>n0, it is determined that it is an abnormal moment. There are two cases for abnormal moment: the first case is noise value (noise influence), and the second case is pollution value (air pollution influence). In order to avoid accidental situation, the method detects N consecutive abnormal moments (N is 5 in this embodiment) before making the following judgment.

[0062] The vector change angle of each two adjacent measurement values of the abnormal moment is calculated, and then the average value of the vector change angle is calculated, as follows:

[0063] As shown in Figure 4 , in the coordinate system, it is assumed that the value of the first abnormal moment is the origin, and the measurement value of the i-th abnormal moment (1 i，i+1 , the measurement value of the i+1-th abnormal moment and the measurement value of the first abnormal moment form a vector, and the two vectors form an angle β 2，3 , β 3，4 , and β 4，5 three angles; the average value of the vector change angle

[0064] If the sensor continuously measures five pollution values, the fluctuation range of the values is small, and the average angle of the vector change is small; when the sensor is affected by noise (air flow, body vibration, sensor hardware measurement fluctuation, etc.), the measurement value fluctuates up and down, and therefore the average vector change angle is large.

[0065] If the average value of the vector change angle is greater than 30°, it is determined that the abnormal moment belongs to the UAV noise influence case; if the average value of the vector change angle is less than or equal to 30°, it is determined that the abnormal moment belongs to the air pollution case.

[0066] Specifically, as shown in Figure 5 (a) and Figure 5 (b), the four measurement values of each graph show consistent change trend, Figure 5 (a) shows an upward trend, Figure 5 (b) shows a stable trend, and the average value of the vector change angle of both is less than 30°, so it is determined to be air pollution. Figure 5 (c) shows an unstable state, and the average value of the vector change angle is greater than 30°, which is usually affected by noise. If this situation occurs, it is determined to be the UAV noise influence case.

[0067] Because of the complexity of the fluctuation, a pollution signal still exists accidentally, so when four pollution signals appear in a certain time (three minutes), the height, longitude and latitude, and pollution value of the point are recorded. When four pollution signals appear, the mean value Value_mean of the pollution value is calculated, and the ratio of the mean value Value_mean to the baseline b0 of the pollution background value of the chemical industry park is calculated, and the calculation formula is as follows:

[0068]

[0069] The pollution level Pollution represents the air pollution condition at the height, and the pollution level can be set according to the gas type: light pollution, moderate pollution, and serious pollution. The above steps can gradually measure the pollution conditions at different heights, and realize the judgment of the pollution level.

[0070] The above-described embodiments are only a preferred scheme of the present application, and are not intended to limit the present application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present application. Therefore, any technical scheme obtained by equivalent replacement or equivalent transformation falls within the protection scope of the present application.

Claims

1. An airborne air pollution detection method based on harmonic vector variation, characterized in that: This air pollution detection method utilizes an unmanned aerial vehicle (UAV) device and a host computer to determine the gas pollution status in the test area. The UAV device includes an aircraft module, a sensor module, and a communication module provided on the aircraft module. The information collected by the sensor module is transmitted to the host computer via the communication module. The specific steps for performing air pollution detection in the host computer are as follows: S1: Construct a harmonic initialization model; the aircraft is at a specified altitude L, with the sensor measurements obtained as the dependent variable and time as the independent variable, and the cosine function amplitude and sine function amplitude in the harmonic initialization model are fitted and solved; S2: Substitute the cosine function amplitude, sine function amplitude and pollution background value baseline of the area to be measured obtained in step S1 into the harmonic initialization model to construct a harmonic model of the area to be measured; S3: After the aircraft reaches the target altitude H, the sensor module obtains the current measurement value Y(t); the current time is substituted into the harmonic model of the measured area constructed in step S2 to obtain the predicted value P(t) at the current time; S4: Based on the difference between the current measured value Y(t) and the current predicted value P(t) and the root mean square error of the harmonic model at the previous moment, an abnormality judgment indication is obtained; if the abnormality judgment indication is less than the abnormality threshold, the current moment is determined to be a normal moment; if the abnormality judgment indication is greater than the abnormality threshold, the current moment is determined to be an abnormal moment; S5: When N abnormal moments are detected continuously, perform vector change angle pollution judgment; calculate the vector change angle of every two adjacent measurement values ​​at the abnormal moment, and then calculate the average value of the vector change angle; if the average value of the vector change angle is greater than 30°, it is judged that the abnormal moment is affected by drone noise; if the average value of the vector change angle is ≤30°, it is judged that the abnormal moment is affected by air pollution.

2. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The harmonic initialization model is as follows: Where: P(t)′ is the sensor measurement value at the current moment, b0 is the pollution background value baseline of the measured area; a1 is the amplitude of the cosine function; b1 is the amplitude of the sine function; T is the vibration period of the cosine function or sine function.

3. The airborne air pollution detection method based on harmonic vector variation according to claim 2, characterized in that: The pollution background value baseline b0 of the test area is measured as follows: the aircraft is located in the test area, and the sensor module is turned on for preheating when not in flight; after the sensor module reaches a stable measurement state, data is collected for T0 minutes and the data obtained by the sensor module is transmitted to the host computer through the communication module; the average of the data obtained by the sensor module within T0 minutes is the pollution background value baseline b0 of the test area; The fitting and solving process of the cosine function amplitude a1 and the sine function amplitude b1 in the harmonic initialization model is as follows: the aircraft is located in the test area, flies to a specified altitude L and maintains it, turns on the sensor module to collect data for T1 minutes and transmits the sensor measurement value P(t)′ to the host computer through the communication module; uses the sensor measurement value P(t)′ as the dependent variable and time as the independent variable to fit and solve the cosine function amplitude a1 and the sine function amplitude b1 in the harmonic initialization model.

4. The airborne air pollution detection method based on harmonic vector variation according to claim 3, characterized in that: The designated height L=3 m; the acquisition time T0 and T1 are both 5 minutes.

5. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The harmonic model of the area to be measured is as follows: Where: P(t) is the predicted value at the current moment; t is the current moment; the target altitude of the aircraft is H>3m; b0 is the baseline of the pollution background value of the area to be measured; a1 is the amplitude of the cosine function; b1 is the amplitude of the sine function; T is the vibration period of the cosine function or the sine function.

6. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The abnormality judgment indication number n in step S4 is calculated as follows: Where t is the current moment, t-1 is the previous moment, delta(t) represents the difference between the measured value Y(t) at the current moment and the predicted value P(t) at the current moment, and RMSE(t-1) represents the root mean square error of the harmonic model at the previous moment.

7. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The abnormal threshold n0 in step S4 is set to 4, which means that the absolute value of the difference between the predicted value and the sensor value is greater than four times the root mean square error.

8. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: When five abnormal moments are detected consecutively in step S5, vector change angle contamination judgment is performed.

9. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The vector change angle pollution judgment in step S5 is specifically as follows: On the coordinate system, assume that the outlier at the first moment is the origin. The measured value at the \(i\)th abnormal moment and the measured value at the first abnormal moment form a vector, where \(1 < i < 5\). The measured value at the \((i + 1)\)th abnormal moment and the measured value at the first abnormal moment form a vector, and the two vectors form an angle \(\beta\). And so on; the average value of the vector change angle i,i+1 , and so on; the average value of the vector change angle where \(N\) is the number of abnormal moments.

10. The airborne air pollution detection method based on harmonic vector variation according to claim 1, characterized in that: The sensor module is a carbon dioxide monitoring sensor, a sulfur dioxide monitoring sensor or a nitrogen dioxide monitoring sensor.

Images