61 results about "Cloud radar" patented technology

The invention discloses an airplane ice accretion monitoring method fusing millimeter wave cloud radar and ground-based microwave radiometer data, which comprises the following steps: step 1, taking temperature profile and relative humidity profile data monitored by a ground-based microwave radiometer in a test area in real time as input quantities to obtain airplane ice accretion index values corresponding to all profile data; step 2, obtaining a reflectivity Z profile of liquid water particles in a test area and a total liquid water value LWP on an observation path; step 3, carrying out inversion to obtain a liquid water content LWC profile in the cloud on the observation path; step 4, taking an overlapped area of an area of which the temperature profile data of the foundation microwave radiometer is less than 0 DEG C and an area of which the inverted liquid water content data of the cloud is greater than 0, a supercooled water area in the cloud is formed; and step 5, analyzing whether the ice accumulation area in the step 1 and the supercooled water area in the step 4 are overlapped or not, if yes, representing that ice accumulation occurs in the air, and if not, representing that no ice accumulation occurs in the air. According to the method disclosed by the invention, a millimeter wave cloud radar and ground-based microwave radiometer active and passive combined detection technology is adopted, the total amount of liquid water and a water vapor profile in the airport airspace are detected, the distribution of supercooled liquid water in cloud in the airport airspace is inverted, and the distribution of a potential icing area of an aircraft is monitored in real time.

The invention discloses a composite precipitation cloud layer identification method based on a laser radar-cloud radar, and the method comprises the steps: obtaining the observation data of the laser radar and the cloud radar, and carrying out the data matching of the observation data; processing a signal of the laser radar in a height interval of 18-20 km to obtain a preprocessed laser radar signal; performing laser radar cloud boundary identification based on a VED algorithm, and endowing different laser radar cloud scores according to the signal attenuation rate and the signal gradient; removing earth surface clutters by using a threshold method based on observation of spectral width and Doppler velocity; carrying out cloud radar significant echo recognition based on a threshold method, and determining the cloud score of the cloud radar according to different reflectivity factor signal intensities; determining a light rain area and a heavy rain area according to the Doppler velocity; cloud layer boundary information and a hairy rain area are obtained through cloud layer recognition result fusion, and cloud layer recognition and rainfall area recognition are completed; according to the invention, the accuracy of weather prediction is improved.

The invention discloses a wall surface identification method and device and a storage medium, and relates to the technical field of sensor positioning, and the method comprises the steps: carrying out the fitting processing of a laser radar point cloud collected by a laser point cloud radar, and obtaining a plurality of fitting line segments; screening a target line segment which is greater than a first length threshold and has the longest length from the plurality of fitting line segments; projecting target point clouds associated with the position of the target line segment in wall point clouds and obstacle point clouds collected by the depth camera on the target line segment, and determining at least one target wall surface line segment; and if the length of the target wall surface line segment is greater than a line segment threshold value, outputting the target wall surface line segment under a robot coordinate system to be used by a welting program. According to the invention, after laser point cloud filtering fitting, depth camera wall surface obstacle point cloud projection calibration is fused, the technical problem that the wall surface is difficult to effectively identify is solved, and the wall attaching precision and obstacle avoidance reliability of the cleaning robot in a complex scene are improved.

The invention discloses a visibility inversion method based on millimeter wave cloud radar and laser radar data fusion, and relates to the technical field of meteorology, and the method comprises the steps: guiding a millimeter wave cloud radar and a laser radar to carry out data correction and equivalent relation construction through employing a fog drop particle size distribution function measured by a fog drop spectrometer in a sample scene in advance; and then applying the correction coefficient and the equivalent function relationship to a target scene to perform multi-source data correction and fusion on the millimeter wave cloud radar and the laser radar and invert visibility. The method can fuse complementary advantages of the laser radar and the millimeter wave cloud radar; the introduction of the fogdrop particle size distribution function effectively avoids the distribution hypothesis of fogdrop shapes in the traditional model, can weaken the seasonal drift of the traditional empirical coefficient, significantly improves the accuracy and reliability of visibility inversion, and achieves the high-precision visibility monitoring under various weather conditions from clear sky to dense fog.

The invention relates to the technical field of radar reflectivity factor calibration, and discloses a terahertz cloud measurement radar calibration method based on channel perception, which comprises the following steps: S1, acquiring two types of input data; s2, constructing a multi-scale channel model; s3, constructing a distance error estimation model; s4, constructing a dynamic time correction model; and S5, constructing a calibration module. According to the method, a high-precision physical channel model is constructed to independently predict path attenuation of signals, the concepts of gain deviation and path attenuation factors are put forward, the influence of thermal noise attenuated by radar on radar waves in the atmospheric environment is reflected respectively, and a Kalman filtering algorithm is used to estimate the path attenuation of the signals. And filtering errors generated in the time direction, and finally, calibrating the radar reflectivity factor by using the channel data by means of a calibration formula.

The embodiment of the invention provides a spaceborne radar and ground-based millimeter wave cloud radar reflectivity factor uniformity evaluation method, which is applied to the technical field of atmospheric detection. The method comprises the following steps: acquiring reflectivity factor data of a spaceborne radar and a ground-based millimeter wave cloud radar; establishing reflectivity factor space-time matching data of the spaceborne radar and the ground-based millimeter wave cloud radar according to the reflectivity factor data by taking vertical direction scanning as a reference; preprocessing the time-space matching data of the reflectivity factors; and based on a preset uniformity evaluation algorithm, an evaluation result is calculated according to the preprocessed reflectivity factor space-time matching data. In this way, errors between original data can be better revealed, and a good foundation is laid for product inversion in the next step.

The invention discloses a cloud phase fine identification and liquid water and ice water quantitative remote sensing device, which comprises a split-phase cloud water quantitative detection laser radar, a novel temperature and humidity remote sensing system and a millimeter wave cloud radar, and is composed of a laser emission module, a telescope receiving module, a microwave radiation module, three laser radar detection modules and a digital signal processing module. The system is composed of a central control processing module and a terminal server. The invention further discloses a remote sensing method of the quantitative remote sensing device. All-day detection and quantitative inversion of liquid water, ice water, water vapor, aerosol and cloud are achieved through the split-phase cloud water quantitative detection laser radar. A novel temperature and humidity remote sensing system is adopted to realize real-time fine detection of a temperature and humidity profile in the cloud. According to the cloud phase state fine identification and liquid water and ice water quantitative remote sensing device and method provided by the invention, the problems that the quantitative detection precision of the liquid water and ice water content in the cloud is insufficient and the mixed phase cloud identification is difficult in the prior art are solved.

The invention provides a cargo volume estimation method and device, electronic equipment and a storage medium, and relates to the technical field of logistics, and the method comprises the steps: obtaining a structured light 3D point cloud, a radar point cloud and an image of a to-be-estimated cargo; performing registration fusion processing on the structured light 3D point cloud, the radar point cloud and the image to obtain fusion data; determining a target type of the to-be-estimated cargo based on the fusion data; and inputting the fusion data into a target volume estimation model corresponding to the target type to obtain the volume of the to-be-estimated cargo. According to the invention, after the fusion data of the to-be-estimated cargo is obtained, the target type of the to-be-estimated cargo is determined based on the fusion data, and then the fusion data is input into the target volume estimation model corresponding to the target type for volume estimation; and the target volume estimation model corresponding to the target type is obtained by training the sample fusion data of the first sample cargo corresponding to the target type, so that the accuracy of cargo volume estimation can be improved.

The embodiment of the invention discloses a cloud shape classification method based on cloud image and millimeter wave cloud radar data fusion, and the method comprises the steps: constructing a visual sub-network, and obtaining a visual feature based on a training cloud image; constructing a millimeter-wave cloud radar sub-network, and obtaining an index token based on a millimeter-wave cloud radar observation value and an index feature after learning a long-distance dependency relationship; constructing a multi-modal fusion sub-network, and obtaining a fusion feature and a final cloud feature after the features of the two are fused; constructing a loss calculation module, and optimizing a cloud classification model composed of the sub-networks by using a loss value to obtain an optimal cloud classification model; and obtaining a cloud shape classification result by using the optimal cloud shape classification model. According to the method, the advantages of the Transform module are fully utilized, the appearance information and the internal structure of the cloud are learned, finally, the features of the Transform module and the internal structure are fused to represent the cloud, and the classification accuracy of the cloud shape is further improved.

The invention relates to a verification method and system for ground-based millimeter wave cloud radar data, and belongs to the technical field of meteorological analysis. The method comprises the following steps: establishing a clear sky echo identification and filtering method, and generating ground-based millimeter wave cloud radar base data after quality control; classifying the ground-based millimeter wave cloud radar base data after quality control into cloud data and rainfall data; the Cloud Sat satellite and the ground-based millimeter wave cloud radar after quality control have reflectivity factor matching and evaluation in cloud data; and matching and evaluating reflectivity factors in GPM satellite and quality-controlled ground-based millimeter wave cloud radar rainfall data. According to the invention, a ground-based millimeter wave cloud radar quality control and classification method is established, a high-quality refined ground-based millimeter wave cloud radar database can be generated, and a good basis is provided for ground-based millimeter wave cloud radar data application; a matching and evaluation method of spaceborne radar and ground-based millimeter wave cloud radar reflectivity factors under different conditions is established, and reference is provided for ground-based millimeter wave cloud radar calibration correction and satellite-ground radar combined application.

For example, a processor may be configured to process point cloud (PC) radar information comprising radar detection information of a plurality of possible detections, wherein radar detection information corresponding to a possible detection of the plurality of possible detections comprises information of a plurality of radar attributes of the possible detection, wherein the processor is configured to determine a plurality of validity scores corresponding to the plurality of possible detections based on the radar detection information of the plurality of possible detections, a validity score corresponding to the possible radar detection to indicate whether it is more probable that the possible detection is a valid detection or a False-Alarm (FA) detection, wherein the processor is to output radar target information based on the plurality of validity scores corresponding to the plurality of possible detections.

The invention relates to a cloud detection method based on combination of a laser radar and a microwave radar, and belongs to the field of radio. The detection method comprises the steps that a coordinate system where a laser radar radiation antenna is located is used as a reference, and system pointing angle deviation caused by the installation position of a microwave radar extension set is corrected in the mode that angle tracking parameters are calculated according to laser-microwave radar extension set antenna installation differences, and scanning space angle synchronization is guaranteed; time synchronization is guaranteed by controlling the time of a radar terminal issuing a control signal to a laser radar extension set and a microwave radar extension set, so that radar beams arrive at meteorological particle echo information to be consistent; after the radar terminal collects data, characteristic parameter extraction and data fusion are carried out on the composite observation data of the laser radar and the cloud radar. According to the method, the problem of target angle deviation caused by spatial distribution difference of the laser radar and the microwave radar during composite observation of the laser-microwave radar system is solved, and the consistency of observed targets is improved.

This invention relates to a cloud height identification method integrating cloud radar and laser wind radar, belonging to the field of radio. The identification method includes: S1, obtaining the cloud base height and cloud top height by inverting the reflectivity product from the cloud radar in top-view mode; S2, obtaining the cloud base height and cloud top height by inverting the signal-to-noise ratio product from the laser wind radar in DBS mode; S3, performing time matching on the output products of the cloud radar and the laser wind radar, and then fusing the results obtained from the cloud radar and the laser wind radar to obtain the final cloud height identification result. This invention complements the characteristics of the two devices. The laser wind radar can compensate for the detection limitations of millimeter-wave radar in thin clouds and distinguish between cloud layers and aerosol layers, while the cloud radar can supplement the inability of laser wind radar to penetrate thick cloud layers, achieving accurate detection of cloud height across the entire cloud area.

The embodiment of the application discloses a kind of multi-modal cloud classification methods based on visual language model, the method comprises: constructing text generation module, generates the cloud description text containing high-level semantic information;Cloud multi-modal feature extraction and fusion module is constructed, cloud image, cloud description text and millimeter wave cloud radar observation data are carried out feature coding and adaptive fusion, obtain multi-modal feature representation;Classifier module is constructed, and the cloud shape classification result of training cloud image is obtained;Total loss function is constructed, and the multi-modal cloud shape classification model consisting of cloud multi-modal feature extraction and fusion module and classifier module is optimized;Test cloud image, and the cloud description text and millimeter wave cloud radar observation data corresponding to test cloud image are input into multi-modal cloud shape classification model and carry out cloud shape classification.The application is realized by introducing high-level semantic cloud description and designing cloud multi-modal adapter, effective fusion of image, radar and text information, improve the accuracy and robustness of cloud shape classification.

The present application discloses a method and device for identifying and predicting micro-rainfall weather processes, which belongs to the field of environmental monitoring technology. The method collects cloud data from Ka-band cloud radar, precipitation data from X-band rain radar, and data from automatic weather stations. A nonlinear model is established, and the data output by the nonlinear model are grouped according to the time axis and radar type. A fitting curve for the micro-rainfall weather process is established to calculate the coefficient of determination and generate eigenvalues, and the prediction result is determined based on the eigenvalues. The output data of the nonlinear model is corrected based on the data from the automatic weather station and the eigenvalues, and the corrected coefficient of determination and eigenvalues ​​are input into the historical data model library as prediction conditions. The judgment results under the same conditions are screened out and compared with the prediction results. If the comparison result is valid, an alarm information push is generated. The method and device for identifying and predicting micro-rainfall weather processes provided by the present application realize the full-process observation and prediction of water vapor-cloud-precipitation, significantly reducing the false alarm rate.

The invention discloses a method for inverting warm cloud microphysical parameters based on a ground-based millimeter wave cloud radar and a microwave radiometer, and the method comprises the steps: 1, obtaining the observation data of the millimeter wave cloud radar, and obtaining a whole-layer liquid water path and a temperature gallery line through the microwave radiometer; 2, identifying warm cloud in millimeter wave cloud radar observation by using a temperature gallery line T; 3, constructing an initial liquid water content profile based on the reflectivity factor and the particle concentration in the warm cloud layer, and calculating an initial liquid water path of the warm cloud layer; 4, the whole-layer liquid water path obtained by the microwave radiometer is used as whole-layer constraint, the initial particle concentration and the initial liquid water content at the vertical height in the warm cloud layer are calibrated according to the proportion, the adjusted warm cloud layer liquid water content and the adjusted warm cloud layer inner particle concentration are obtained, and a physical boundary constraint interval is applied to the particle concentration; and 5, calculating the effective radius of the particles in the warm cloud layer according to the adjusted liquid water content of the warm cloud layer and the particle concentration in the warm cloud layer. According to the method, high-precision warm cloud microphysical parameter vertical distribution can be obtained.

The present invention designs a method for intelligent recognition of non-meteorological targets based on deep learning, and identifies non-meteorological targets existing in millimeter-wave cloud radar data; specifically, by designing a millimeter-wave cloud radar data annotation algorithm and a data iteration algorithm, the corresponding labels are quickly obtained; a data expansion method for expanding millimeter-wave cloud radar data is designed to obtain more diverse training data, and feature extraction and feature restoration are designed to improve the recognition performance of non-meteorological targets; and an excellent recognition model is designed using the encoder-decoder concept. Through the above set of processes, the accuracy of intelligent recognition of non-meteorological targets of millimeter-wave cloud radar is improved. Furthermore, the present invention enhances prompts for specific labels by constructing a prompt enhancement module, thereby improving the sensitivity of the model to specific labels, reducing the computational complexity of the intelligent recognition model, and improving the generalization ability.

The application discloses a millimeter wave cloud radar echo classification method and system based on physical guidance deep learning, comprising: constructing a multi-source fusion dataset, aligning, standardizing and sliding window sampling preprocessing millimeter wave cloud radar base data and co-sited ground observation data; extracting thermodynamic state information from the ground observation data through a physical constraint gating module, generating a clear sky echo probability graph to adaptively weight the radar initial features; based on an encoder-decoder backbone network, the weighted radar features are jointly convolved and extracted, and an ECH regression branch is used to fuse the thermodynamic features and the radar features to predict the effective clutter height; based on the predicted effective clutter height sequence, an adaptive height filtering region is constructed, and morphological opening operation and intersection operation of dynamic height constraint are used to realize fine echo classification post-processing, so that the height balance of meteorological echo retention rate and clear sky echo suppression rate is realized, and the data quality of the millimeter wave cloud radar is effectively improved.

For example, a processor may be configured to process point cloud (PC) radar information comprising radar detection information of a plurality of possible detections, wherein radar detection information corresponding to a possible detection of the plurality of possible detections comprises information of a plurality of radar attributes of the possible detection, wherein the processor is configured to determine a plurality of validity scores corresponding to the plurality of possible detections based on the radar detection information of the plurality of possible detections, a validity score corresponding to the possible radar detection to indicate whether it is more probable that the possible detection is a valid detection or a False-Alarm (FA) detection, wherein the processor is to output radar target information based on the plurality of validity scores corresponding to the plurality of possible detections.

The invention relates to an object recognition method, a storage medium, a program product, electronic equipment and a vehicle. The object identification method comprises the following steps: eliminating an object in a target area in a first image to obtain a second image; and identifying a target object in the target area according to the first image and the second image. According to the embodiment of the invention, the method does not depend on a laser point cloud radar or a plurality of cameras, can achieve the recognition of the obstacle even through the image collected by one monocular camera, does not need the hardware equipment for image collection, and can adapt to different hardware configurations.

This application provides a millimeter-wave cloud radar observation method and related apparatus based on multi-source data. A fused dataset is formed by real-time acquisition and preprocessing of at least one of the following: weather radar mosaic data, meteorological satellite data, ground visibility observation data, and co-located or adjacent aerosol lidar data. A preset rule set is used to determine that the current weather scene at the cloud radar station includes at least weather echo approaching, weather echo receding, local weather processes, fog, and non-precipitating clouds. Based on the identified weather scene type, the corresponding combined scanning strategy (composed of a basic scanning mode consisting of vertical top-view scanning, distance-height scanning, and three-dimensional volume scanning, and a fine scanning mode consisting of local planar horizontal scanning and layered planar horizontal scanning, combined according to preset logic) is called from a preset strategy mapping table and executed. During execution, the parameters of the fine scanning mode are dynamically adjusted based on the position and intensity information of the real-time radar echo or changes in the detection mode.

The utility model discloses a shelter system for artificially influencing weather based on an unmanned aerial vehicle. A command and control cabin, an observation cabin and a transportation and storage cabin are arranged in the square cabin, and a skylight corresponding to the observation cabin is arranged on the top wall of the square cabin; the observation cabin is provided with a cloud radar, the transportation and storage cabin is provided with an unmanned aerial vehicle, and the command and control cabin is provided with accessory equipment matched with the cloud radar and the unmanned aerial vehicle; a first cabin door is arranged on the side wall of the square cabin, corresponds to the transportation and storage cabin and is used for entering and exiting of the unmanned aerial vehicle; the two ends of the lifting mechanism are connected to the bottom wall in the square cabin and the bottom of the cloud radar respectively, and the lifting mechanism is configured to lift the cloud radar out of the square cabin during detection operation and store the cloud radar into the observation cabin during non-operation. The unmanned aerial vehicle and the cloud radar are ingeniously integrated in the high-maneuverability square cabin, the site turnover requirement caused by regional seasonal changes is met, the efficiency and effect of artificial influence weather operation are improved, and the operation cost is effectively reduced.

The invention relates to a method for realizing cloud shape automatic identification based on a ground-based millimeter wave cloud radar, and the method comprises the following steps: S10, carrying out the quality control of collected cloud radar data, and removing abnormal echo data; s20, performing cloud boundary identification based on the data after quality control to obtain data sets of different cloud blocks; s30, calculating a cloud feature quantity by using the data sets of different cloud blocks; and S40, cloud identification is carried out based on the characteristic quantity. According to the method, the cloud radar is used for detecting data, through observation data analysis, quality control, cloud boundary identification and cloud characteristic quantity calculation, the cloud characteristic quantity is used as a basis, a threshold value method and a K-nearest neighbor machine learning method are used, automatic cloud identification is achieved, automatic cloud observation of a meteorological system is improved, and the method is suitable for automatic cloud observation of the meteorological major.

The invention relates to a cloud height identification method based on fusion of a cloud radar and a laser wind finding radar, and belongs to the field of radio, and the identification method comprises the steps: S1, obtaining the cloud bottom height and the cloud top height through the inversion of a reflectivity product of a cloud radar top-to-top mode; s2, performing inversion by using a signal-to-noise ratio product in a laser wind finding radar DBS mode to obtain a cloud bottom height and a cloud top height; and S3, performing time matching on output products of the cloud radar and the laser wind finding radar, and performing product fusion on results obtained by the cloud radar and the laser wind finding radar to obtain a final cloud height identification result. According to the invention, the characteristics of the two devices are complementary, the laser wind-finding radar can make up for the detection defect of the millimeter wave radar on the thin cloud and distinguish the cloud layer from the aerosol layer, the cloud radar can make up for the defect that the laser wind-finding radar cannot penetrate through the deep cloud layer, and the cloud height global precise detection is realized.

The application discloses a kind of spaceborne cloud radar beam conical spiral scanning forward-looking and rearview observation fast registration method, belong to meteorological radar signal processing and spaceborne observation registration technical field.The method includes: according to satellite orbit parameter and scanning geometry, the ground intersection of each pulse beam footprint is calculated, and footprint point set is constructed;According to the interval of scanning circle, the potential overlapping observation is traversed, and the candidate point pair is generated in combination with spatial distance gate;Matching cost function is constructed with footprint distance as main item and fusion observation consistency constraint, and effective registration point pair is obtained under one-to-one correspondence constraint;Registration point pair obtained by different scanning circle interval and different distance gate is accumulated and deduplicated, and global registration point set is formed.The present application makes full use of multiple overlapping observations of spiral trajectory, significantly increases the same sample, improves the reliability and precision of registration under the condition of high-speed spaceborne platform, and can provide high-quality input for subsequent three-dimensional observation data fusion, data quality control and data assimilation.

The invention discloses a cloud recognition method based on multiple observation devices, which is characterized in that a weather phenomenon video intelligent observation instrument, an aerosol laser radar, a millimeter wave cloud radar and other observation devices are used for comprehensive judgment to finally obtain a recognition result, and comprises the following steps: acquiring a sky image from the weather phenomenon video intelligent observation instrument, and judging whether cloud exists or not; the cloud base height measured by the laser radar is combined to judge whether the cloud is high cloud or low cloud; non-cloud echoes and noise echoes are filtered through a quality control algorithm, and a reliable cloud radar product is obtained; the boundary of the cloud is identified through the cloud radar product, so that the cloud information such as the cloud bottom height, the cloud top height, the cloud layer thickness and the cloud layer number is obtained. The method is small in calculation amount, high in recognition rate and wide in application range, and overcomes the defects that in the prior art, manual work is relied on, and the processing speed is low.

This application provides a cloud measurement system and method based on a wave glider. By integrating a laser cloud-measuring radar with a marine meteorological observation instrument onto a wave glider platform with ultra-long endurance, autonomous navigation, and virtual mooring capabilities, a sea-based mobile cloud observation system is constructed. This system overcomes the inherent shortcomings of existing satellite remote sensing observations, such as discontinuity and lack of vertical structure information, as well as the high cost, poor maneuverability, and limited coverage of ship and buoy observations. It achieves long-term, continuous, and mobile observation of cloud systems over vast ocean areas and can acquire high-precision cloud base height, vertical structure, and conventional meteorological elements from the sea surface, significantly improving the spatiotemporal continuity and data quality of marine meteorological observation.

This invention discloses a method and system for identifying layered cloud melting layers based on cloud radar, belonging to the field of cloud radar identification technology. The method includes: acquiring airborne cloud radar observation data; resampling the airborne cloud radar observation data to a unified geographic grid; for each sample, extracting continuous time periods with radar bright band characteristics and performing time averaging to obtain a comprehensive average reflectivity factor profile; calculating the vertical gradient of the comprehensive average reflectivity factor profile; and identifying the effective melting layer based on the vertical gradient. This invention automatically identifies the height of the top and bottom of the melting layer by analyzing the vertical gradient changes of reflectivity factor and vertical velocity, exhibiting high accuracy and strong adaptability.

The invention discloses a snowfall inversion method based on a millimeter wave cloud radar and a microwave radiometer, and relates to the technical field of radar snowfall inversion, and the method comprises the following steps: S1, data input: millimeter wave cloud radar data and microwave radiometer data are input; s2, multi-source equipment data space-time dimension matching is carried out; s3, phase state classification of extremely cold particles, and elimination of non-snowfall phase state interference; s4, inverting the snowfall rates of different height layers; s5, the accumulated snowfall of the whole-layer snowfall rate is obtained through time integration; and S6, outputting a result. According to the invention, the complementary advantages of high temporal-spatial resolution of the millimeter wave cloud radar and abundant atmospheric environment parameters of the microwave radiometer are brought into full play, the inherent defect of single equipment inversion is solved, and the precision and reliability of snowfall inversion in the extremely cold environment are greatly improved; and a data support is provided for practical applications such as meteorological monitoring and early warning in extremely cold areas, and prevention and reduction of snowstorm disasters.

The present invention provides a cloud physics parameter inversion method based on a convolutional neural network, comprising the following steps: S1, calculating the scattering characteristics of cloud particles in the terahertz frequency band, combining the particle spectrum distribution, simulating and calculating the functional relationship between the terahertz radar equivalent reflectivity factor and the cloud physics characteristic parameters, and establishing a forward physical model; S2, preprocessing the measured data and microphysical parameter products of the terahertz cloud radar to obtain a training database; S3, establishing the structure of the convolutional neural network, using the forward physical model as a constraint, and adaptively adjusting the learning rate based on the deviation between the predicted value and the actual value of the radar echo; S4, repeatedly training the convolutional neural network using the training database, and obtaining the cloud physics characteristic parameters by inverting the terahertz radar equivalent reflectivity factor. This method improves the inversion accuracy of cloud physics parameters, eliminates the reliance on empirical values, and is an algorithm with stronger generalization and adaptive learning.