Smoke plume boundary identification method, electronic equipment and readable storage medium
By processing the backscattered data of Doppler wind measuring lidar, deleting blind spots and deviation data, and calculating the gradient of standard deviation ratio, the problem of inaccurate identification of smoke plume boundaries in the existing technology is solved, and more accurate smoke plume is achieved Boundary identification.
Patent Information
- Application Number
- CN202411185347.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-11-22
AI Technical Summary
In the prior art, it is difficult to accurately determine the boundary of the plume by using the change of the backscattering intensity of the Doppler wind measurement lidar, and the existing algorithms are prone to misidentify the maximum value inside the plume as the boundary, resulting in inaccurate identification of the plume boundary.
By obtaining the backscattering data of Doppler wind measuring lidar, deleting blind spot data and data with radial velocity deviation greater than the threshold, calculate the standard deviation of the backscattering intensity at three to five distance gates and the starting distance gate to correspond to the standard deviation of the backscattering intensity at the three to five distance gates. The gradient of the ratio of the sum of the standard deviations of the backscattering intensity of the distance gate, determines the boundary position of the plume.
It reduces noise sensitivity, more accurately recognizes the end boundary of the plume, and improves the recognition accuracy of the plume boundary.
Smart Images

Figure CN119001665A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of remote sensing, and in particular to a smoke plume boundary recognition method, electronic equipment and a readable storage medium. Background Art
[0002] The occurrence of forest fires not only causes a large amount of forest resources to be wasted, but also has a serious impact on the ecological environment. Therefore, forest fire detection and identification is an important research field. Smoke plume refers to the flow state of aerosols (such as smoke, dust and other tiny particles) in the atmosphere under certain conditions. As an important signal of fire, smoke plume has the characteristics of fast diffusion speed and wide spread range. It can reflect the spread direction and degree of fire. It has shown broad application prospects in the fields of fire detection, air quality assessment and fire behavior analysis. Therefore, taking appropriate measures to monitor forest fires is an important task in the current forest fire prevention and control.
[0003] Since the smoke plume produced by forest fires is visible in the early stages of a fire, early detection of smoke plumes is an effective way to avoid losses caused by fires, thereby achieving early detection and early warning of fires and minimizing the damage caused by fires to forests. Traditional forest monitoring methods include ground patrols, lookout patrols, aerial patrols, and video surveillance, but they have the disadvantages of high labor costs, untimely confirmation of fire points, delayed discovery time, and low efficiency.
[0004] With the rapid development of ground remote sensing instruments and technologies, scanning lidar, as the most appropriate tool for ground monitoring of wildfire plume dynamics and different distances from wildfires, can continuously monitor the polluted atmosphere near forest fires and provide real-time information on the properties of aerosols in the atmosphere, plume height, and the spatiotemporal changes in the direction and speed of plume movement. It is sensitive to changes in the concentration of aerosol particles generated by fires, has high accuracy and resolution, and relevant personnel can operate safely at a location far away from the burning area. It can play an important role in the field of forest fire plume monitoring. Using scanning lidar to monitor and prevent forest fires is an efficient and economical means currently and in the future.
[0005] When the lidar scans the smoke plume, if the corresponding backscatter signal intensity increases or decreases sharply at a certain position, it can be determined that this position is the boundary or boundary layer of the smoke plume. This is because factors such as the aerosol concentration and size in the smoke plume will affect the intensity of the backscatter signal, thereby reflecting the boundary characteristics of the aerosol plume. The backscatter intensity refers to the signal reflected back by the target object after the lidar emits a laser beam to the target object. Therefore, using the change in the backscatter signal intensity to determine the boundary or boundary layer of the smoke plume is a common aerosol measurement method.
[0006] In order to study the smoke plume produced by forest fires and describe its changing characteristics such as height and thickness, the backscatter intensity of the Doppler wind lidar is used to determine the near boundary of the smoke plume. However, since the backscatter intensity varies greatly in each scan of the lidar, it is difficult to determine the boundary of the smoke plume by using a specific threshold, resulting in inaccurate identification of the smoke plume boundary.
[0007] In addition, the prior art also uses an algorithm to determine the boundary of the smoke plume by taking the maximum gradient of the integrated ratio of the backscatter intensity from the current range gate to the backscatter intensity from the starting range gate to the current range gate. However, when this algorithm is applied, the maximum value may be found at a location that is not necessarily the edge boundary of the smoke plume, but at a point inside the smoke plume, which also leads to inaccurate identification of the smoke plume boundary. Summary of the invention
[0008] In view of the above-mentioned deficiencies in the prior art, a smoke plume boundary recognition method, an electronic device and a readable storage medium provided by the present invention solve the problem of inaccurate smoke plume boundary recognition in the prior art.
[0009] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:
[0010] A method for identifying a smoke plume boundary is provided, which comprises the following steps:
[0011] S1, obtain backscatter data of Doppler wind lidar;
[0012] S2, deleting the blind area data in the backscatter data, and deleting the data that exceeds the range gate and whose radial velocity deviation is greater than the deviation threshold as invalid data, to obtain the preprocessed data;
[0013] S3, obtaining the standard deviation of the backscatter intensity at three to five range gates r based on the preprocessed data, recorded as STD(r);
[0014] S4. Obtain the sum of the standard deviations of the backscattering intensity from the starting range gate to the corresponding range gate r according to the preprocessed data, denoted as STD(r min -r);
[0015] S5. Calculate STD(r) and STD(r min -r), and the position of the maximum gradient value is taken as the boundary position of the smoke plume scanned by the Doppler wind laser radar.
[0016] Furthermore, the blind area data in the backscatter data is the first 60 meters of the backscatter data.
[0017] Furthermore, the deviation threshold is ±7 meters per second.
[0018] Furthermore, STD(r) and STD(r min -r) is expressed as:
[0019]
[0020] Where D STD(r) Indicates STD(r) and STD(r min -r) ratio gradient; It means to find the gradient.
[0021] An electronic device is provided, comprising:
[0022] a memory storing executable instructions; and
[0023] The processor is configured to execute the executable instructions in the memory to implement the smoke plume boundary identification method.
[0024] A readable storage medium is provided, on which executable instructions are stored, and when the executable instructions are executed by a processor, a method for identifying a smoke plume boundary is implemented.
[0025] The beneficial effects of the present invention are as follows: the present invention first removes the blind area data within the first range gate, and then performs secondary elimination on the remaining data through the range gate parameters and the radial velocity deviation threshold, so that the retained data is effective backscatter data; the position of the smoke plume boundary is determined by calculating the gradient of the ratio of the backscatter intensity standard deviation at three to five range gates r in the effective backscatter data to the sum of the backscatter intensity standard deviation from the starting range gate to the corresponding range gate r, which can reduce noise sensitivity and more accurately identify the end boundary of the smoke plume. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 is a schematic diagram of the process of this method;
[0027] Figure 2 This is the smoke plume boundary recognition effect diagram of the existing gradient algorithm;
[0028] Figure 3 This is the smoke plume boundary recognition effect diagram of this method;
[0029] Figure 4 Figure 2 shows the backscatter intensity as a function of the range gate. (a) shows the backscatter intensity as a function of the range gate, and (b) shows the front edge (black) and the end edge (red) of the plume as a function of the range gate. STD(r) ;
[0030] Figure 5 A scatter plot comparing the real range gate at the front end of the plume and the range gate identified by the algorithm;
[0031] Figure 6This is a scatter plot comparing the true range gate at the end of the plume and the range gate identified by the algorithm. DETAILED DESCRIPTION
[0032] The specific implementation modes of the present invention are described below so that those skilled in the art can understand the present invention. However, it should be clear that the present invention is not limited to the scope of the specific implementation modes. For those of ordinary skill in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the attached claims, these changes are obvious, and all inventions and creations utilizing the concept of the present invention are protected.
[0033] like Figure 1 As shown, the smoke plume boundary recognition method includes the following steps:
[0034] S1, obtain backscatter data of Doppler wind lidar;
[0035] S2, deleting the blind area data in the backscatter data, and deleting the data that exceeds the range gate and whose radial velocity deviation is greater than the deviation threshold as invalid data, to obtain the preprocessed data;
[0036] S3, obtaining the standard deviation of the backscatter intensity at three to five range gates r based on the preprocessed data, recorded as STD(r);
[0037] S4. Obtain the sum of the standard deviations of the backscattering intensity from the starting range gate to the corresponding range gate r according to the preprocessed data, denoted as STD(r min -r);
[0038] S5. Calculate STD(r) and STD(r min -r), and the position of the maximum gradient value is taken as the boundary position of the smoke plume scanned by the Doppler wind laser radar.
[0039] The electronic device comprises:
[0040] a memory storing executable instructions; and
[0041] The processor is configured to execute the executable instructions in the memory to implement the smoke plume boundary identification method.
[0042] A readable storage medium stores executable instructions, which implement a smoke plume boundary recognition method when executed by a processor.
[0043] In the specific implementation process, the range gate is one of the parameters of the Doppler wind laser radar. Since the Doppler wind laser radar cannot measure within the first range gate, this results in a "blind zone" in the first 60m (the range gate is 60 meters). In order to correct the instrument error, this embodiment deletes any data within the first 60m to eliminate any potential damaged data. In addition, the validity of the backscatter data inside and outside the plume is determined by checking the returned Doppler radial velocity. At a certain range gate, the radial velocity fluctuates due to signal attenuation. Therefore, this embodiment determines that all data that exceeds the range gate and has a radial velocity deviation of ±7m / s are invalid and deleted. Since the radial velocity data exceeding this threshold is considered unrealistic, its corresponding backscatter intensity data is also considered invalid and is therefore not used for subsequent analysis, so that any boundary determined by this method is within the area of valid data.
[0044] STD(r) and STD(r min -r) is expressed as:
[0045]
[0046] Where D STD(r) Indicates STD(r) and STD(r min -r) ratio gradient; It means to find the gradient.
[0047] In one embodiment of the present invention, both the existing gradient algorithm and the present method are applied to the backscatter data of the laser radar stare scan, such as Figure 2 and Figure 3 As shown in the figure (black scattered points represent the boundary points of the front end of the plume, and red scattered points represent the boundary points of the end of the plume), although the existing gradient algorithm can roughly determine the boundary position of the plume, the gradient algorithm is poor in recognizing weak backscattering intensity, resulting in poor recognition of the near-end plume boundary and sensitivity to random noise at the far end. On the contrary, compared with the gradient algorithm, this method can better find the boundary position of the plume, and this method is less sensitive to random noise at the far end, and can more accurately identify the boundary of the end of the plume.
[0048] In another embodiment of the present invention, Charland et al. pointed out that when the laser radar scans the aerosol plume, if the corresponding backscattered signal intensity increases or decreases sharply at a certain position, it can be determined that this position is the boundary or boundary layer of the aerosol plume. In order to verify the effectiveness of this method in identifying the front and end boundaries of the plume, therefore, based on the theory of Charland et al., this embodiment carries out the verification work of this method in plume boundary identification. Figure 4 It can be seen from the figure that when the backscattering intensity reaches its peak, D can be identified by this method. STD(r)The two peaks appear at the front and end boundaries of the backscattering intensity peak (plume), where the red dotted lines indicate the positions.
[0049] To further verify the reliability of this method in identifying the front and end boundaries of the smoke plume, this example continues to perform statistical analysis on the lidar observation data obtained from the forest fires that occurred in British Columbia, Canada from August 18 to August 23, 2018. The position where the backscatter signal intensity increases sharply at a certain position is taken as the real range gate of the front end, and the position where it decreases sharply is taken as the real range gate of the end, which are compared and verified with the front range gate and the end range gate of the smoke plume identified by this method, and finally 8706 sample points are obtained.
[0050] like Figure 5 and Figure 6 As shown (the red line indicates the reference line of Y=X), the correlation coefficient R between the front range gate identified by this method and the real front range gate is 0.94, and the root mean square error RMSE is 4.58; the correlation coefficient R between the terminal range gate identified by the algorithm and the real terminal range gate is 0.92, and the root mean square error RMSE is 5.29. It can be seen that the correlation coefficients of this method in identifying the front and terminal boundaries of the smoke plume are both above 0.9, with a small deviation from the actual situation, showing good results and also illustrating the effectiveness and reliability of this method.
[0051] In summary, compared with the existing gradient algorithm, the present invention can better find the boundary position of the smoke plume, and the present invention has reduced sensitivity to random noise at the far end, and can more accurately identify the end boundary of the smoke plume. The present invention is not only applicable to PPI laser radar scanning data, but also to Stare laser radar scanning data to determine the boundary of the smoke plume, providing a new idea and method for identifying the boundary of the forest fire smoke plume.
Claims
1. A method for identifying a smoke plume boundary, characterized in that: The following steps are involved: S1, obtain backscatter data of Doppler wind lidar; S2, deleting the blind area data in the backscatter data, and deleting the data that exceeds the range gate and whose radial velocity deviation is greater than the deviation threshold as invalid data, to obtain the preprocessed data; S3, obtaining the standard deviation of the backscatter intensity at three to five range gates r based on the preprocessed data, recorded as STD(r); S4. Obtain the sum of the standard deviations of the backscattering intensity from the starting range gate to the corresponding range gate r according to the preprocessed data, denoted as STD(r min -r); S5. Calculate STD(r) and STD(r min -r), and the position of the maximum gradient value is taken as the boundary position of the smoke plume scanned by the Doppler wind laser radar.
2. The method for identifying smoke plume boundaries according to claim 1, characterized in that: The blind area data in the backscatter data is the first 60 meters of the backscatter data.
3. The method for identifying smoke plume boundaries according to claim 1, characterized in that: The deviation threshold is ±7 meters per second.
4. The method for identifying smoke plume boundaries according to claim 1, characterized in that: STD(r) and STD(r min -r) is expressed as: Where D STD(r) Indicates STD(r) and STD(r min -r) ratio gradient; It means to find the gradient.
5. An electronic device, characterized in that: include: A memory storing executable instructions; as well as A processor is configured to execute the executable instructions in the memory to implement any one of the methods of claims 1 to 4.
6. A readable storage medium having executable instructions stored thereon, characterized in that: When the executable instructions are executed by a processor, the method described in any one of claims 1 to 4 is implemented.
