High-precision precipitation particle detection method and system

Abstract

This invention discloses a high-precision precipitation particle detection method and system. It relates to the field of optical and meteorological measurement equipment technology, solving the problems of poor accuracy and insufficient adaptability in precipitation particle parameter measurement in existing technologies. This invention corrects the distortion of precipitation particle slice contours caused by these movements through horizontal and vertical velocities, thereby accurately identifying the shape, size, flatness, three-dimensional orientation, and type of precipitation particles during their descent. Based on the degree of change in the upper and lower sets of precipitation particle contours, the rotation rate of the precipitation particles is estimated. An artificial intelligence recognition algorithm is used to train and analyze the precipitation particle contours to identify the type of precipitation particles. A photoelectric detection unit detects the difference in scattering intensity signals of precipitation particles to two light sources, and a threshold is used to determine the phase state of the precipitation particles, thus improving the ability to identify precipitation particle types.

Description

Technical Field

[0001] This invention relates to the field of optical and meteorological measurement equipment technology, specifically to a high-precision precipitation particle detection method and system. Background Technology

[0002] Precipitation particle characteristics are the result of the combined effects of cloud microphysics and atmospheric motion, and are of great significance in fields such as cloud precipitation physics, weather modification, aviation meteorological support, flood forecasting and early warning, and weather radar. The characteristic parameters of precipitation particles mainly include shape, aggregation state, size, flattening, orientation, descent velocity, and type. Traditional measurement methods, such as the grid method, rapid photography method, immersion method, flour ball method, and filter paper staining method, are unsuitable for measuring and analyzing large amounts of data due to their large workload and inability to automatically complete measurement and classification.

[0003] To address this, instruments such as laser raindrop spectrometers and two-dimensional imaging raindrop spectrometers have emerged on the market, capable of automatically recording liquid and solid precipitation particles such as raindrops, snowflakes, and hail without human intervention. Among these, laser raindrop spectrometers calculate the equivalent size and falling velocity of precipitation particles by measuring the intensity and duration of obstruction of their emitted laser beam by precipitation particles. However, they cannot obtain the contour information of the precipitation particles, nor can they distinguish between multiple precipitation particles falling simultaneously.

[0004] Two-dimensional imaging raindrop spectrometers mainly employ two high-speed line scan cameras that are orthogonal in the horizontal direction and maintain a preset parallel distance in the vertical direction, along with two background light sources. They detect the contours of precipitation particles from two orthogonal directions and calculate the falling velocity by using the time it takes for the precipitation particles to pass between the two cameras, thus achieving the measurement of the characteristic parameters of precipitation particles. However, there are still many issues that need to be improved.

[0005] The existing two-dimensional imaging raindrop spectrometer has two cameras that are not on the same plane. When precipitation particles have horizontal velocities, it is impossible to obtain the accurate distance between the precipitation particles and the two cameras. In this case, the estimation of the horizontal size of the precipitation particles will have a large deviation.

[0006] Existing two-dimensional imaging raindrop spectrometers cannot accurately detect the horizontal velocity and orientation of precipitation particles. In operation, they assume that precipitation particles are symmetrical and fall without tilt, estimating horizontal displacement and thus horizontal velocity by measuring the displacement between the centers of the first and last scan lines of the precipitation particle using a camera. However, many actual precipitation particles are asymmetrical (such as larger raindrops, graupel, and snow / ice polymers), and their falling process involves orientation (tilt angle) and horizontal motion.

[0007] When precipitation particles are in horizontal motion, the outline of the precipitation particles captured by the camera will produce an apparent tilt angle. If the precipitation particles have an actual tilt angle, it is impossible to distinguish between the actual tilt angle and the apparent tilt angle generated by the horizontal motion. At this time, the particle size and velocity estimated by the equipment are no longer valid.

[0008] For precipitation particles that are tilted but not horizontally moving, the device increases the apparent height and decreases the apparent width, thereby increasing the measured flatness, and the measured precipitation particle profile will be severely deformed.

[0009] Current two-dimensional imaging raindrop spectrometers lack level and azimuth detection. The orientation of precipitation particles is based on the horizontal plane and azimuth, but it is difficult to ensure the accuracy of level and azimuth at all times during equipment installation and use. Even small changes in level and azimuth will seriously affect the accuracy of orientation data.

[0010] In addition, if the equipment is tilted, its actual sampling area will deviate from the nominal sampling area, thereby increasing the measurement error of cumulative rainfall and rainfall intensity.

[0011] Existing two-dimensional imaging raindrop spectrometers use a fixed threshold to identify precipitation particles. When the background light source of the equipment decays due to aging, changes in ambient temperature and humidity, or changes in atmospheric visibility, the threshold cannot be dynamically adjusted, the sensitivity of the equipment will decrease, and precipitation particles may be missed or the equipment may even fail to function properly, increasing the difficulty of equipment maintenance.

[0012] Existing two-dimensional imaging raindrop spectrometers lack the function of detecting the phase state of precipitation particles, which is one of the most important parameters of precipitation particles.

[0013] Existing two-dimensional imaging raindrop spectrometers cannot detect the rotation rate of precipitation particles.

[0014] Existing two-dimensional imaging raindrop spectrometers are large in size, consume a lot of power, and are difficult to maintain. They are mostly used in remote areas where power supply and maintenance are difficult. Therefore, it is imperative to improve their efficiency, reduce power consumption, reduce losses, and extend their service life. Summary of the Invention

[0015] To address the shortcomings of existing technologies, this invention provides a high-precision precipitation particle detection method and system, which solves the problems of poor measurement accuracy and insufficient adaptability of precipitation particle parameters (such as shape, size, flatness, orientation, falling velocity, and type) in existing technologies.

[0016] To achieve the above objectives, the present invention provides the following technical solution: a high-precision precipitation particle detection system, comprising a housing, the upper and lower ends of which are sealed by an upper cover and a lower cover; The housing contains a main control circuit, two or more sets of particle detection units, an environmental detection unit, and a heating and defrosting unit. The main control circuit is equipped with a gyroscope, electronic compass, satellite positioning and time synchronization unit, and is used to provide signal processing and control circuitry. The particle detection unit consists of two mutually orthogonal high-speed linear scanning camera units located in the same plane and two background light sources in different bands (e.g., infrared and visible light). One camera unit uses visible light as the background light source, and the other camera unit uses infrared light as the background light source. The two camera units and their respective corresponding light sources are synchronously modulated and then work asynchronously to avoid mutual interference between the two orthogonal directions. The two camera units and their respective corresponding light sources are synchronously modulated and then work asynchronously to avoid mutual interference between the two orthogonal directions. The two or more sets of particle detection units are placed horizontally and stacked, and maintain a preset distance in the vertical direction. Of the two or more particle detection units, at least one of the particle detection units is equipped with a photoelectric detection unit. An environmental detection unit is installed inside or outside the housing; A heating and defrosting unit is installed in the particle detection window area inside the housing; The housing is fitted with a windshield.

[0017] A high-precision precipitation particle detection method includes the following steps: A three-dimensional XYZ coordinate system is established using the detection system. Two orthogonal linear scanning cameras, A and B, are positioned at height Z in the precipitation particle detection unit. Camera A's measurement plane is perpendicular to the X-axis (yz-plane imaging), its linear array is along the Y-direction, and its scanning direction is along the Z-direction. Camera B's measurement plane is perpendicular to the Y-axis (xz-plane imaging), its linear array is along the X-direction, and its scanning direction is along the Z-direction. When precipitation particles fall through the precipitation particle detection unit, the contours of the precipitation particles in two orthogonal directions are obtained by two high-speed line scan camera units. The location of the precipitation particles first obtained by camera A is the distance from the precipitation particles to camera B, and vice versa. Then, the horizontal dimensions of the precipitation particles are calculated using the actual distance between the precipitation particles and the camera, and the camera parameters. Specifically, the width and length of the precipitation particles are calculated using Formula 1, as follows: Formula 1: W R =L D ×L C ×S C Among them, W R L represents the actual horizontal size of precipitation particles. D Indicates the shooting distance, L C S represents the pixel size of precipitation particles in a camera image.C This indicates the pixel size of the camera sensor.

[0018] This invention assumes that the vertical interval (distance in the Z direction) between the two sets of horizontal orthogonal linear scanning cameras of precipitation particle detection units is ΔZ. The specific method for calculating the horizontal velocity (Vx, Vy) and vertical velocity (Vz) of precipitation particles is as follows: The first precipitation particle detection unit (upper group) consists of two orthogonal linear scanning cameras, A1 and B1, positioned at height Z1. Camera A1's measurement plane is perpendicular to the X-axis (yz-plane imaging), with the linear array along the Y-direction and the scanning direction along the Z-direction. Camera B1's measurement plane is perpendicular to the Y-axis (xz-plane imaging), with the linear array along the X-direction and the scanning direction along the Z-direction.

[0019] The second set of precipitation particle detection units (lower set): two orthogonal line scan cameras A2 and B2, arranged at height Z2, where Z2 = Z1 - ΔZ (the lower set is below the upper set). A2 is parallel to A1, and B2 is parallel to B1. The corresponding cameras in both sets have the same (aligned) projection position on the XY plane, ensuring that the vertical baseline is purely in the Z direction.

[0020] When the system is working, each camera performs a high-speed line scan. As a raindrop passes through the measurement plane of each camera group, a spatiotemporal image is captured in each camera. Each camera group can determine the three-dimensional position of the raindrop within its measurement plane. Specifically: For the precipitation particle detection unit in the upper group: When camera A1 sees a raindrop on a certain scan line (time t), it obtains the (y, z) coordinates of the raindrop at that moment, but the x-coordinate is given by the fixed XA1 of the camera's measurement plane in space (known calibration value). When camera B1 sees the raindrop at the same moment (synchronization time), it obtains the (x, z) coordinates, and the y-coordinate is fixed at YB1 (known calibration value). However, the z-coordinates seen by the two cameras should be consistent (if at the same moment), so that the three-dimensional points of the raindrop can be matched to obtain the precise spatial position (x1, y1, z1) of the raindrop at time t1.

[0021] Similarly, for the next set of precipitation particle detection units, the positions (x2, y2, z2) are matched at a later time t2.

[0022] Because the upper and lower groups are a fixed distance ΔZ in the Z direction, the time difference between the raindrops passing through the upper and lower groups is Δt = t2 - t1 (which can be accurately obtained from the high-speed scanning timestamp).

[0023] Because raindrops move continuously, each camera sees a trajectory (coordinates on a line at multiple moments). We need to find the spatial point corresponding to a certain moment t1 in the upper group and the corresponding spatial point at moment t2 for the same raindrop in the lower group. The determination of the same precipitation particle can be achieved by ensuring consistency in velocity direction and ensuring reasonable time and position differences between the upper and lower groups. Algorithmically, (t2, t1) pairings can be found using feature matching or minimizing projection error. The horizontal and vertical velocities are obtained using the following formulas: Vx=(x2-x1) / Δt, Vy=(y2-y1) / Δt, Vz=(z2-z1) / Δt Furthermore, based on the velocity measurements of the two sets of orthogonal line-scan cameras mentioned above, this invention obtains the shape, size, and three-dimensional orientation of precipitation particles (raindrops, snowflakes, etc.), and utilizes the multi-angle, multi-slice imaging capabilities of the system for three-dimensional reconstruction. The principles and methods are explained in detail below.

[0024] Whenever a precipitation particle passes through the plane of the precipitation particle detection unit (taking the A1 camera as an example), the camera sees a two-dimensional projection (YZ direction) of the precipitation particle on that plane (fixed X=XA1). However, the camera sensor is a one-dimensional linear array, which can only obtain the intensity distribution of a line along the Y direction on that plane (this line corresponds to a certain Z coordinate). Since the particles move in the Z direction (Vz≠0) and the scanning time is very fast, different scan lines (different times) correspond to different Z coordinate positions of the precipitation particles on that plane, thus obtaining a slice (Y direction contour) along the Z direction on the plane XA1, that is, obtaining a two-dimensional projection image (YZ image) of the precipitation particles on the plane XA1.

[0025] Within a single measurement plane, the camera obtains the projection of precipitation particles onto that plane, rather than a direct three-dimensional voxel. For non-spherical particles, the shape of this projection changes with the particle's orientation.

[0026] By combining slices from two orthogonal cameras—namely, the projection of precipitation particles onto plane XA1 (YZ profile) obtained by camera A1, and the projection of precipitation particles onto plane YB1 (XZ profile) obtained by camera B1—two orthogonal projections of precipitation particles can be obtained. These two projections are mutually perpendicular viewpoints, and they image the same particle at the same time (or almost simultaneously).

[0027] The two cameras in the lower group obtained two orthogonal projection sequences of precipitation particles at another altitude after a time Δt.

[0028] Using known velocities, slices at different times are projected onto the particle coordinate system (compensation motion), so that all slices correspond to different cross sections when the particle is at rest.

[0029] Each slice is the line integral of precipitation particles on that plane (fixed X or fixed Y). Multiple slices at different Z positions are equivalent to projection data (parallel beam projection) from that viewpoint. Parallel beam projection data from two orthogonal viewpoints can be used for 3D tomographic reconstruction (similar to CT). Since there are only two angles, the reconstruction may be blurry, but by combining it with a prior shape model (such as rotational symmetry, ellipsoids, etc.), the 3D shape can be optimized.

[0030] Using the reconstructed three-dimensional shape of precipitation particles, the size, flattening, orientation, and equivalent volume of the precipitation particles are calculated to determine the shape category of the precipitation particles (e.g., spherical, ellipsoidal, irregular).

[0031] The three-dimensional orientation of precipitation particles is usually represented by three Euler angles. Alternatively, the orientation of the particles in three-dimensional space can be directly inferred from the principal axes of the contours of two orthogonal projections.

[0032] If the contours of the upper and lower groups change significantly, the rotation rate can also be estimated. For example, within the time difference Δt between the upper and lower groups, precipitation particles may rotate or oscillate. If we assume that the orientation of the precipitation particles remains unchanged within Δt, then the upper and lower groups are equivalent to imaging the same precipitation particle from the same viewpoint at two different Z positions. This provides information on two cross-sections of the particle along the direction of motion (Z), which helps determine shape changes along the Z direction (such as whether the particle is symmetrical). If the projected contours of the upper and lower groups change, it can be inferred that the particle is rotating or oscillating. This provides higher-dimensional information for the study of precipitation particles.

[0033] During installation or use, the equipment may be tilted (not installed horizontally) or have azimuth errors (e.g., not aligned with north). This invention uses a gyroscope and electronic compass to detect and automatically record the system's tilt and azimuth angles in real time. Based on these angles, the three-dimensional coordinate system is rotated to a horizontal XY plane, with the X-axis pointing north and the Z-axis pointing upwards and perpendicular to the ground. The position and orientation of precipitation particle profiles in space are then remapped using the rotated coordinate system.

[0034] The profiles of precipitation particles are trained and analyzed using artificial recognition algorithms to identify the types of precipitation particles (such as rain, snow, hail, sleet, etc.).

[0035] The photoelectric detection unit in the particle detection unit detects the scattering intensity signal of precipitation particles to the light source. By calculating and analyzing this signal, the phase state of the precipitation particles is determined according to the threshold.

[0036] The camera image mode can be either black and white or color. When the camera image mode is in color, it can capture real images of precipitation particles, making it easier to identify precipitation types and further improving the ability to identify precipitation particle types.

[0037] The photoelectric detection unit in the particle detection unit detects the intensity signal of atmospheric scattering of light source, calculates atmospheric visibility, and controls the power intensity of light source or dynamically adjusts the detection threshold of camera based on atmospheric visibility data, thereby improving the sensitivity of precipitation particle detection.

[0038] By using a windshield to reduce the horizontal velocity of precipitation particles in the detection area, the effective detection data of the equipment is improved during periods of strong winds and precipitation.

[0039] By installing a heating defrosting unit in the detection window area of ​​the equipment, rain, dew, frost, snow, ice and other substances are prevented from adhering, ensuring unobstructed access to the detection window area; The system acquires data such as ambient temperature, humidity, air pressure, wind speed, and wind direction through an environmental detection unit, calculates the dew point temperature of the air, determines the presence of frost and dew conditions, automatically activates the heating and defrosting unit, and adjusts the heating power according to the ambient temperature, i.e., the lower the temperature, the greater the heating power; when the environment no longer has frost and dew conditions, the system automatically or by receiving a command from the host computer turns off the heating.

[0040] The tilt angle of the system is detected in real time by using gyroscopes and compasses. When the tilt angle exceeds a given threshold, on-site maintenance is carried out in a timely manner, and the data is cleaned before it is used later.

[0041] The system obtains the geographical location, altitude, and time of the detection system through satellite positioning and time synchronization units, and performs regular time synchronization on the equipment.

[0042] By setting three operating modes—low frequency, high frequency, and sleep trigger—in the main control circuit hardware and software, the following three operating modes are provided for users to flexibly choose from: (1) Continuous high frequency working mode: In this mode, regardless of whether there is precipitation, the equipment will scan the sampling space at full load and capture precipitation particles to the maximum extent, but it has high requirements for the power consumption of the equipment.

[0043] (2) Normal low-frequency working mode: Under normal conditions, the topmost camera and its corresponding light source are kept in a low-speed scanning state to detect whether precipitation particles are present. Once precipitation particles are detected, the high-frequency working mode is automatically activated. After a preset time has elapsed since the precipitation ended, the low-frequency working mode is automatically activated again. This mode can effectively reduce the power consumption and component wear of the equipment.

[0044] (3) Normal sleep trigger working mode: The photoelectric detection unit in the particle detection unit continuously detects the atmospheric state, while other functions are in sleep mode. When the photoelectric detection unit detects precipitation particles, it automatically wakes up the high-frequency working mode. After the precipitation ends for a preset time, it automatically enters the sleep trigger mode again. This mode can greatly improve the efficiency of equipment use, reduce power consumption, reduce losses and extend service life.

[0045] This invention provides a high-precision method and system for detecting precipitation particles. Compared with existing technologies, it has the following advantages: The invention's precipitation particle detection unit consists of two mutually orthogonal high-speed line-scan camera units that are on the same plane. Each unit can provide the other with the accurate distance of the precipitation particles relative to itself. Based on this distance and camera parameters, the size of the precipitation particles can be accurately calculated, thus improving the detection accuracy of particle size-related parameters.

[0046] This invention can detect the position evolution of precipitation particles in three-dimensional space during their fall by using two or more sets of precipitation particle detection units, and based on this information, the horizontal velocity (Vx, Vy) and vertical velocity (Vz) of the precipitation particles can be accurately calculated. The present invention further corrects the distortion of the precipitation particle slice profile caused by these movements by using horizontal and vertical velocities, thereby accurately identifying the shape, size, flatness, three-dimensional orientation and type of precipitation particles during their fall.

[0047] Based on the degree of change in the outlines of the upper and lower sets of precipitation particles, this invention can also estimate the rotation rate of precipitation particles, a function that is not available in commercially available products.

[0048] This invention improves the signal-to-noise ratio of the detection by synchronously modulating the two camera units and their corresponding light sources in the precipitation particle detection unit and then operating them asynchronously, thus avoiding mutual interference between the two orthogonal light sources.

[0049] This invention uses a gyroscope and electronic compass to detect and automatically record the tilt and azimuth of the system in real time. Based on the tilt and azimuth, the coordinate system of the observation system can be rotated to the correct position, and the parameters of each precipitation particle can be recalculated, thereby improving the detection accuracy of parameters such as horizontal velocity, vertical velocity, three-dimensional orientation, and rotation rate of precipitation particles.

[0050] This invention can revise the actual sampling area of ​​the system according to the system tilt angle, thereby improving the measurement accuracy of cumulative rainfall and rainfall intensity.

[0051] When an abnormality in tilt or azimuth is detected, this invention can promptly carry out on-site maintenance, and can also clean and control the data when using it later.

[0052] This invention uses artificial intelligence recognition algorithms to train and analyze the profiles of precipitation particles to identify the types of precipitation particles.

[0053] This invention improves the ability to identify precipitation particle types by detecting the difference in scattering intensity signals of precipitation particles to two light sources through a photoelectric detection unit and determining the phase state of precipitation particles based on a threshold.

[0054] The camera image mode of this invention can be either black and white or color. When the camera image mode is in color, it can capture real precipitation particle images, which facilitates the identification of precipitation types and further improves the ability to identify precipitation particle types.

[0055] This invention uses a photoelectric detection unit to detect the intensity signals of atmospheric scattering of light at different wavelengths, thereby obtaining atmospheric visibility signals. Based on the visibility signals, the power intensity of the light source can be controlled or the camera detection threshold can be dynamically adjusted, further improving the accuracy of precipitation particle detection.

[0056] This invention reduces the horizontal velocity of precipitation particles in the detection area by using a windbreak, thereby improving the effective detection data of the equipment during periods of strong winds and precipitation.

[0057] This invention acquires environmental parameters through an environmental detection unit, determines the presence of frost and dew conditions, and automatically activates the heating and defrosting unit. When the environment no longer provides frost and dew conditions, it automatically or by receiving a command from the host computer shuts down the heating. This intermittent heating function significantly reduces the equipment's energy consumption, making it more suitable for the low-power requirements of field observation.

[0058] When frost conditions are detected, this invention adjusts the heating power according to the ambient temperature; that is, the lower the temperature, the greater the heating power. This can better solve the problems of dew, frost, and snow removal in extremely cold environments, and can also effectively save the energy consumption of the equipment, making it more suitable for the low power consumption requirements of field observation.

[0059] This invention acquires the geographical location, altitude, and time of the detection system through satellite positioning and time synchronization units, and performs regular time synchronization on the equipment. This data is crucial for subsequent research and application of precipitation particle variation patterns.

[0060] This invention improves the efficiency of the device, reduces power consumption, minimizes losses, and extends its service life by setting three working modes—low frequency, high frequency, and sleep trigger—in the main control circuit hardware and software, providing users with three flexible working modes to choose from. Attached Figure Description

[0061] Figure 1 This is a three-dimensional schematic diagram of the detection system of the present invention; Figure 2 This is a top view schematic diagram of the detection system of the present invention; Figure 3 This is a bottom view schematic diagram of the detection system of the present invention; Figure 4 This is a front view schematic diagram of the detection system of the present invention; Figure 5 This is a rear view schematic diagram of the detection system of the present invention; Figure 6 This is a cross-sectional view A of the detection system of the present invention; Figure 7 This is a cross-sectional view B of the detection system of the present invention; Figure 8 This is a cross-sectional view C of the detection system of the present invention; Figure 9 This is a cross-sectional view D of the detection system of the present invention; Figure 10 Figure A shows the internal structure of the detection system of this invention. Figure 11 Figure B shows the internal structure of the detection system of this invention. Figure 12 Figure C shows the internal structure of the detection system of this invention. Figure 13 This is an internal diagram (D) of the detection system of the present invention; Figure 14 Figure E shows the internal structure of the detection system of this invention. Figure 15 Figure F shows the internal structure of the detection system of this invention. Reference numerals: 1. Windshield; 2. Housing; 3. Heating and defrosting unit; 4. Main control circuit; 5. Camera; 6. Light source; 7. Photoelectric detection unit; 8. Environmental detection unit. Detailed Implementation

[0062] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0063] First Embodiment Please see Figure 1 This application provides a high-precision precipitation particle detection method, including the following steps: When precipitation particles fall through the precipitation particle detection unit, the projected contours of the precipitation particles in two orthogonal directions can be detected by two orthogonal high-speed line scan camera units. The vertical coordinates of the precipitation particles seen by the two cameras should be consistent (if at the same moment), so that the three-dimensional points of the raindrops can be matched to obtain the precise spatial position of the precipitation particles at this moment.

[0064] Similarly, when precipitation particles fall through the lower set of precipitation particle detection units, the projected profiles in two orthogonal directions are obtained again, and the precise spatial position of the precipitation particles at this moment is obtained. Because the upper and lower groups are a fixed distance ΔZ in the vertical direction, the time difference Δt between the raindrops passing through the upper and lower groups can be accurately obtained from the high-speed scanning timestamp. The horizontal and vertical velocities of precipitation particles can be accurately calculated using spatial location and time difference.

[0065] Furthermore, based on the velocity measurements of the two sets of orthogonal line scan cameras mentioned above, the system's multi-angle and multi-slice imaging capabilities are used to perform three-dimensional reconstruction, obtaining the shape, size, and three-dimensional orientation of precipitation particles (raindrops, snowflakes, etc.). By analyzing images from two sets of orthogonal cameras, detection capabilities that other systems cannot achieve can be provided, such as the rotation rate of precipitation particles.

[0066] In the particle detection unit, each high-speed line scan camera unit can provide the other with the accurate distance of the precipitation particles relative to itself, and calculate the accurate size of the precipitation particles based on this distance, thereby improving the detection accuracy of particle size-related parameters. In the particle detection unit, two high-speed scanning camera units are synchronously modulated with their respective corresponding light sources. During operation, the two high-speed scanning camera units work asynchronously to avoid mutual interference between the two orthogonal directions of detection, thereby improving the signal-to-noise ratio of the detection. By using a gyroscope and electronic compass, the system tilt angle and azimuth angle are detected and recorded in real time. The accurate position of the precipitation particle profile in the three-dimensional coordinate system can be precisely revised based on the tilt angle and azimuth angle. The position and orientation of the precipitation particle profile in space can be remapped based on the rotated coordinate system. The profiles of precipitation particles are trained and analyzed using artificial recognition algorithms to identify the types of precipitation particles collected (such as rain, snow, hail, sleet, etc.). The photoelectric detection unit in the particle detection unit detects the scattering intensity signals of precipitation particles to two light sources. By analyzing these signals, the phase state of the precipitation particles can be further determined, thus improving the ability to identify the type of precipitation particles. Furthermore, the camera image mode can be either black and white or color. When the camera image mode is color, white light is used as the background light source, and a foreground light source is added to one side of the camera. In this mode, precipitation types can be identified by capturing real precipitation particle images, further improving the ability to identify precipitation particle types.

[0067] By detecting the atmospheric light scattering intensity signal through the photoelectric detection unit in the particle detection unit, the atmospheric visibility signal can be obtained. Based on the visibility signal, the power intensity of the light source can be controlled or the camera detection threshold can be dynamically adjusted, further improving the accuracy of precipitation particle detection. By using windshields to reduce the horizontal velocity of precipitation particles in the detection area, the effective detection data of the equipment is improved during periods of strong winds and precipitation. By installing a heating defrosting unit in the detection window area of ​​the equipment, rain, dew, frost, snow, ice and other substances are prevented from adhering, ensuring unobstructed access to the detection window area; The system acquires data such as ambient temperature, humidity, air pressure, wind speed, and wind direction through an environmental detection unit, calculates the dew point temperature of the air, determines the presence of frost and dew conditions, automatically activates the heating and defrosting unit, and adjusts the heating power according to the ambient temperature, i.e., the lower the temperature, the greater the heating power; when the environment no longer has frost and dew conditions, the system automatically or by receiving a command from the host computer turns off the heating. By monitoring the tilt angle of the system in real time, when the tilt angle exceeds a given threshold, on-site maintenance is carried out in a timely manner, and the data is cleaned before it is used later. The system obtains its geographical location, altitude, and time through satellite positioning and time synchronization units, and performs regular time synchronization on the equipment. By setting three operating modes—low frequency, high frequency, and sleep trigger—in the main control circuit hardware and software, the following three operating modes are provided for users to flexibly choose from: (1) Continuous high frequency working mode: In this mode, regardless of whether there is precipitation, the equipment will scan the sampling space at full load and capture precipitation particles to the maximum extent, but it has high requirements for the power consumption of the equipment.

[0068] (2) Normal low-frequency working mode: Under normal conditions, the topmost camera and its corresponding light source are kept in a low-speed scanning state to detect whether precipitation particles are present. Once precipitation particles are detected, the high-frequency working mode is automatically activated. After a preset time has elapsed since the precipitation ended, the low-frequency working mode is automatically activated again. This mode can effectively reduce the power consumption and component wear of the equipment.

[0069] (3) Normal sleep trigger working mode: The photoelectric detection unit in the particle detection unit continuously detects the atmospheric state, while other functions are in sleep mode. When the photoelectric detection unit detects precipitation particles, it automatically wakes up the high-frequency working mode. After the precipitation ends for a preset time, it automatically enters the sleep trigger mode again. This mode can greatly improve the efficiency of equipment use, reduce power consumption, reduce losses and extend service life.

[0070] Considering cost-effectiveness, the detection system of the present invention includes at least two sets of particle detection units, one above the other, and each set of particle detection units includes at least two mutually orthogonal camera units.

[0071] Furthermore, the above is a preferred embodiment of the present invention. In actual implementation, the stacking of two or more sets of particle detection units is not limited to the vertical direction, and the camera units in each set of precipitation particle detection units are not limited to the orthogonal direction. More camera units and light sources can be added at different angles in the precipitation particle detection units to obtain slice projections of precipitation particles from more angles, thereby enhancing the accuracy of three-dimensional reconstruction of precipitation particles and obtaining more accurate precipitation particle parameters.

[0072] Second Embodiment High-precision precipitation particle detection system, such as Figures 1 to 15 As shown, it includes a windshield 1. The windshield has a large opening at the top and a small opening at the bottom. Its shape is similar to a funnel, which helps to guide the wind blowing from all directions on the outer surface and reduce the impact on the detection area. Its side has many through holes, which help to slow down the wind speed. The housing 2 consists of upper and lower covers. The outer surface of the upper cover has a structure for fixing a windshield, and its top is sloping to prevent the accumulation of rain, snow, and hail. It also contains a heating and defrosting unit to prevent frost buildup on the camera, light source, and other particle detection devices and related optical path structures. The interior contains a fixed structure for the particle detection units (camera, light source, photoelectric detection unit, etc.), and each component of the particle detection unit has corresponding openings. The hollow structure serves as the particle observation area, running vertically to facilitate the descent of rain, snow, and other particles. Its function is to provide structural support and a suitable detection structure for the entire particle detection system, reducing the impact of irrelevant factors on detection accuracy.

[0073] The heating and defrosting unit 3 is a module with heating function, such as a PI, PTC, PET or silicone heating film. It is attached to the inner surface of the housing, such as around the camera, light source and other related particle detection devices, around the related optical path structure and the top inner surface, to prevent frost, ice and snow from forming on the detection unit and optical path.

[0074] The main control circuit 4 serves as the main controller for the camera, light source, photoelectric detection unit, environmental detection unit, etc., controlling the entire timing and process of particle detection, and processing and transmitting the collected data; it can be equipped with onboard gyroscope, electronic compass, satellite positioning and other modules, and has functions such as real-time detection and automatic recording of system tilt angle and azimuth angle, as well as time calibration.

[0075] The camera 5 consists of a lens, a photosensitive circuit board, and a circuit protection housing. A typical particle detection unit consists of two cameras positioned orthogonally. Each camera has a photosensitive circuit to collect signals, and one of the two photosensitive circuit boards can serve as the main control circuit. The camera's circuit protection housing is fixed inside the housing 2, which has an opening for the lens of the camera 5, facing the particle detection area. Its function is to image precipitation particles from multiple angles and in multiple slices, and to process the collected signals or directly transmit them to the main control for processing and 3D reconstruction, thereby obtaining data such as the shape, size, flattening, 3D orientation, falling velocity, rotation rate, and type of precipitation particles.

[0076] The light source 6 can be a point light, line light, or area light source, and its type can be LED, etc. Its placement can also be adjusted according to detection needs, such as placing it in front of or behind the camera. The light source frequency can also be adjusted as needed. The wavelength of the light source can be adjusted according to detection needs; for example, a single light source can have multiple wavelengths. The number of light sources can also be adjusted; multiple cameras can be equipped with one light source, one camera can be equipped with one light source, or one camera can be equipped with multiple wavelength light sources.

[0077] The photoelectric detection unit 7 can be a photon detector or a photoelectric detector, or an array combination of these sensors. It is fixed on the outer shell 2 and placed in the particle scattering area. It can be on the same plane as a group of particle detection units. Its position can be adjusted according to the structure and detection requirements. It can also be placed in multiple positions as needed. Its function is to detect the scattering intensity signal of light of different wavelengths in the particle detection area.

[0078] The environmental detection unit 8 can be equipped with sensor elements such as temperature, humidity, pressure, wind speed, and wind direction, or other environmental measurement elements can be added as needed. It can be fixed on the main control board, photoelectric detection circuit board, or separately fixed inside or outside the housing 2 according to actual detection needs. It can detect the external environment in real time, such as acquiring data on ambient temperature, humidity, air pressure, wind speed, and wind direction, and transmit them to the main control for heating control or other related data processing.

[0079] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0080] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A high-precision precipitation particle detection system, including: The casing, main control circuit, particle detection unit, environmental detection unit, heating and defrosting unit, and windshield are characterized by: The housing contains a main control circuit, two or more sets of particle detection units, an environmental detection unit, and a heating and defrosting unit. A [missing information] is located in the middle of the housing. The main control circuit is equipped with a gyroscope, electronic compass, satellite positioning and time synchronization unit, and is used to provide signal processing and control circuits; The particle detection unit mainly consists of two mutually orthogonal high-speed line scan camera units located on the same plane and two or more background light sources in different bands.

2. The high-precision precipitation particle detection system according to claim 1, characterized in that, Two or more sets of the particle detection units are placed horizontally, stacked, and maintained at a preset distance in the vertical direction.

3. The high-precision precipitation particle detection system according to claim 1, characterized in that, Of the two or more sets of particle detection units, at least one set of particle detection units is equipped with a photoelectric detection unit.

4. The high-precision precipitation particle detection system according to claim 1, characterized in that, The high-speed line scan camera image modes include black and white mode or color mode.

5. The high-precision precipitation particle detection system according to claim 1, characterized in that, The environmental detection unit is installed inside or outside the housing, and the heating defrosting unit is installed inside the housing at the particle detection window area.

6. The high-precision precipitation particle detection system according to claim 1, characterized in that, The main control circuit is equipped with three working modes: low frequency, high frequency, and sleep mode. High-frequency operating mode: The load performs high-frequency scanning sampling of the space to capture precipitation particles to the maximum extent. Low-frequency working mode: The topmost camera and its corresponding light source are kept in a low-speed scanning state to detect whether precipitation particles are present; Sleep mode: The photoelectric detection unit in the particle detection unit continuously detects the atmospheric condition, while other functions are in sleep mode. When the photoelectric detection unit detects precipitation particles, it automatically wakes up the high-frequency working mode. After a preset time when the precipitation ends, it automatically enters the sleep trigger mode again.

7. A high-precision precipitation particle detection method, wherein the method is implemented according to any one of claims 1-6, characterized in that, Includes the following steps: The projected profiles of precipitation particles in two orthogonal directions are detected by two orthogonal high-speed line scan camera units during the falling process. When the precipitation particles fall through the next set of precipitation particle detection units, the projected profiles in two orthogonal directions are obtained again, thus obtaining the precise spatial position of the precipitation particles. By utilizing the multi-angle, multi-slice imaging capability of the detection system, three-dimensional reconstruction is performed to obtain the shape, size, and three-dimensional orientation of precipitation particles. The high-speed line scan camera unit provides the other party with the accurate distance of precipitation particles relative to this unit, and calculates the accurate size of the precipitation particles based on this distance.

8. The high-precision precipitation particle detection method according to claim 7, characterized in that, The gyroscope and electronic compass detect and automatically record the system tilt and azimuth angles in real time. Based on the tilt and azimuth angles, they accurately revise the position of the precipitation particle profile in the three-dimensional coordinate system and remap the position and orientation of the precipitation particle profile in space according to the rotated coordinate system.

9. The high-precision precipitation particle detection method according to claim 7, characterized in that, The profiles of precipitation particles are trained and analyzed using artificial recognition algorithms to identify the types of precipitation particles collected.

10. The high-precision precipitation particle detection method according to claim 7, characterized in that, The photoelectric detection unit in the particle detection unit detects the scattering intensity signals of precipitation particles to two light sources, and analyzes these signals to determine the phase state of the precipitation particles.

Images