144results about How to "Reduce computing power consumption" patented technology

According to the chart visualization implementation method and system based on natural language processing, input texts are analyzed on the basis of a natural language processing model, so that key visualization types are obtained, and charts matched with the visualization types are generated on the basis of the visualization types. The target natural language processing network is a natural language processing network obtained by loading the text learning sample to an initial natural language processing network to be adjusted for adjustment; the text learning sample is composed of a plurality of template implicit representations obtained by extraction of a plurality of text paragraph training templates, and mapping data types and feedback signals corresponding to the template implicit representations, and preamble implicit representations in a text paragraph set are adopted; network adjustment is carried out in a mode of fusion and feedback of preorder text implicit representation and current text implicit representation, and the accuracy of target visual text visualization type recognition is improved.

The invention discloses a non-linear state rapid fitting method for an aircraft, and belongs to the field of aircraft control. The implementation method comprises the following steps: constructing a two-loop attitude pilot and a three-loop attitude pilot, and discretizing to obtain a guidance instruction with a time-delay effect; according to coordinate conversion, forward deriving each aerodynamic parameter of the aircraft and an aircraft overload instruction equation; the method comprises the following steps: calculating all pneumatic data according to part of pneumatic data obtained by a blowing experiment of an aircraft, and fitting a guidance instruction with a balance attack angle, a sideslip angle and a rudder deflection angle of the aircraft to obtain a corresponding interpolation table; all aerodynamic parameters in the flight process of the aircraft are obtained according to an aircraft guidance instruction and an interpolation table in the actual situation, namely, nonlinear state rapid fitting of the aircraft is achieved, and simulation reproduction of the aircraft as a mass point in the whole guidance process is achieved according to the nonlinear state rapid fitting result. And the precision and efficiency of aircraft guidance control are improved.

The invention discloses a charging pile operation state evaluation method and system, and the method comprises the following steps: S1, building a cross-source data integration library based on the operation state sample data of a charging pile, S2, carrying out the layering processing of the cross-source data integration library, S3, carrying out the segmentation processing of a charging pile cross-source data set, and S4, carrying out the evaluation of the operation state of the charging pile. Calculating the fatigue degree of the cross-source data set of the charging pile according to the cross-source data integration library, and marking when the fatigue degree exceeds a fatigue degree threshold value; s5, for the marked charging pile cross-source data set, the sensitivity of the charging pile cross-source data set is calculated according to the cross-source data integration library, and when the sensitivity exceeds a sensitivity threshold value, marking is carried out; s6, the fatigue degrees and the acuity degrees of all the charging piles in the charging pile cross-source data set are calculated independently, when the specific fatigue degree and the specific acuity degree exceed threshold values, marking is carried out, and the evaluation grade is improved; the normal operation state of the charging pile can be ensured to the maximum extent, the monitoring efficiency can be improved, and the computing power consumption is reduced.

The invention relates to the technical field of virtual reality rendering, particularly provides a rendering picture generation method, a controller, a storage medium and augmented reality equipment, and aims to solve the problem of how to provide rendering pictures for multiple wearable equipment to save cost and avoid insufficient computing power when the augmented reality equipment comprises the multiple wearable equipment. In order to achieve the purpose, the virtual scene is rendered through a main renderer of a first processor, the rendering middle frame with the field angle larger than the field angle of the wearable device is obtained, the rendering picture of each wearable device is obtained according to the rendering middle frame and the orientation of each wearable device, the rendering middle frame is obtained only through one-time rendering, and the rendering efficiency is improved. According to the embodiment of the invention, the rendering intermediate frame is adjusted, and the rendering picture of each wearable device is obtained, so that the process of performing multi-view rendering on each wearable device can be avoided, and the computing power consumption is effectively reduced on the premise of ensuring the effect of the rendering picture of each wearable device.

The invention relates to a mobile robot dynamic obstacle real-time detection method and system based on embedded vision. The method comprises the following steps: performing motion compensation on a historical frame image, generating a motion compensation image, performing frame difference analysis on the motion compensation image and a current frame image, performing adaptive threshold segmentation in combination with local texture features, and generating a motion candidate mask; and extracting feature points in the mask region, calculating an optical flow vector and generating RGB image blocks, and performing dynamic attribute verification after fusion to obtain a verified dynamic point set. And further calculating an optical flow direction angle and a fusion speed value, constructing a motion feature vector and clustering, and generating a central position coordinate of the obstacle instance. And through rotation matrix conversion and Kalman filtering processing, a smooth position estimation value and a smooth speed estimation value are obtained, and the relative distance and the dynamic threat degree of the obstacle are calculated. According to the method, through motion compensation, filtering optimization and the like, the real-time performance, precision and motion estimation accuracy of dynamic obstacle detection in an embedded scene are improved.

The embodiment of the invention provides a display device, a head-up display and a vehicle, the display device is applied to the head-up display, the display device emits image light, and the image light is reflected into an eye box through an imaging window corresponding to the head-up display, so that a user can observe a virtual image formed by the image light. The curvature of the imaging window in at least one of the transverse direction and the longitudinal direction is larger than 0 so as to form a first distortion form, the display device comprises a liquid crystal display screen and a backlight module, the backlight module is installed on the back side of the liquid crystal display screen and used for providing backlight for the liquid crystal display screen so that the liquid crystal display screen can emit image light, and the backlight module comprises a plurality of backlight lamps; the brightness of at least one of the plurality of backlights can be independently adjusted according to the virtual image content of the virtual image, the curvature of the shape formed by arranging the plurality of backlights in at least one of the transverse direction and the longitudinal direction of the liquid crystal display screen is greater than 0, and the bending direction of the shape is opposite to the bending direction of the corresponding area in the imaging window.

The invention discloses an image transmission method, device and equipment of a remote operation system and a medium, the method is applied to a local operation end, the remote operation system comprises the local operation end and a remote operation end, and the method comprises the following steps: acquiring an initial ultrasonic image corresponding to a target object at the local operation end, performing feature extraction on the initial ultrasonic image to obtain a target ultrasonic image; determining an ultrasonic sub-image corresponding to the target ultrasonic image based on the knowledge graph, and determining an image value of the ultrasonic sub-image according to the ultrasonic image evaluation dimension; and transmitting the ultrasonic sub-image to a remote operation end according to the image value. Based on the technical scheme, the image is divided into the sub-images of different areas, the image values of the sub-images are determined, and the transmission methods corresponding to the sub-images are determined according to the image values, so that the preferential transmission of the high-value image is ensured, the overall bandwidth occupation and the system load are reduced, and the transmission efficiency is improved. And a key technical support is provided for the accuracy and safety of a remote operation.

The invention provides an intelligent lamp post equipment data processing method and system based on vehicle and road cloud, and the method comprises the steps: loading a picture sample library to a basic image processing network in a network training link; the method comprises the following steps: constructing map unit basic features corresponding to a plurality of map units based on a network, constructing involved connecting lines, and determining connecting line basic features corresponding to the involved connecting lines according to shared debugging parameters of the involved connecting lines and the map unit basic features; besides, a basic graph is constructed, network parameters of a basic image processing network can be updated according to graph unit basic features, recognized object two-tuples and the basic graph corresponding to every two object graph units, and the image processing network used for recognizing the object two-tuples containing the recognized object relation in the traffic image is obtained. According to the invention, the generalization and flexibility of the network can be improved, and the training speed of the network is improved.

The invention relates to the field of in-memory calculation, and provides an in-memory calculation method and device of a floating-point number exponent and a storage medium, the method is applied to an in-memory calculation structure of the floating-point number exponent, the structure comprises a static random access memory (SRAM) row, a first delay chain and a maximum value searching and difference value generating module, the first delay chain is formed by connecting N in-memory computing units (CIMs) in series; according to the in-memory calculation method of the floating-point number index, an input floating-point number index and a weight floating-point number index are divided into a high part and a low part, maximum value searching and difference value generation are carried out on the high part and the low part respectively, and double-edge index summation calculation is carried out on a rising edge and a falling edge of an output pulse. The amount of data needing to be stored and the number of calculation units are reduced, the circuit area of calculation in the floating-point number index memory can be reduced, and the calculation time and the calculation power consumption can be reduced.

The invention provides a power-saving processing method of a wireless communication system, electronic equipment and a storage medium, and relates to the technical field of wireless communication. The method comprises the steps that a main processor determines whether to enter a power-saving mode, and if yes, the main processor pushes a communication protocol stack program to a power-saving control module; the power-saving control module starts a simplified protocol stack program according to the communication protocol stack program and sends starting response information to the main processor; the main processor executes a reset operation according to the starting response information, communicates with the network equipment to complete resource suspension, and sends a power-off request to the power-saving control module; and the power-saving control module controls the main processor to be powered off according to the power-off request, and performs network communication through the switching chip based on a simplified protocol stack program. According to the method, idle power-off of the main processor can be realized, and quick power-on recovery after power-off can be realized by running a simplified protocol stack program in the power-saving control module.

The invention relates to the technical field of composite material structural member design, in particular to a composite material pipe fitting bending rigidity prediction method based on deep learning and a system and a terminal thereof. The composite material pipe fitting bending rigidity prediction method comprises the following steps: testing the bending resistance of a layered composite material pipe fitting by using a finite element model under a cantilever beam condition; establishing a basic data set; processing the basic data set to obtain a standard data set; constructing a neural network model, and training the neural network model by using the standard data set to obtain a pipe bending rigidity prediction model; and outputting the flexural rigidity by using the model. The method comprises the following steps: acquiring data by simulating a composite material pipe fitting, dividing the preprocessed data into layer paving data and pipe fitting geometric data as independent variables and flexural rigidity as dependent variables, making a sample data set, constructing a pipe fitting flexural rigidity prediction model, and calculating the flexural rigidity of the pipe fitting through basic data. And low-cost, rapid and accurate design of the composite material pipe fitting is realized.

The invention provides a channel prediction method and system based on a multi-cycle feature convolutional network. According to the method, context sensing weighted input and a dynamic multi-model pool mechanism are introduced. A terminal device obtains environmental context features in real time, dynamically calculates the weight of each time step in a historical channel state information sequence based on the features, and generates a weighted optimized input sequence. Meanwhile, the network equipment pre-trains and maintains a plurality of multi-period feature convolutional network models specific to different environments, and the network can effectively capture the complex time-varying rules in the period and during the weeks in the channel, so that the accuracy of channel state information prediction is improved. And the terminal equipment immediately selects and activates the optimal sub-model in the model pool for prediction according to the real-time context features, and locally performs increment fine adjustment. According to the invention, through management of the data input end and rapid and adaptive switching of the prediction model end, the precision of channel prediction and the adaptive capacity to environmental sudden change are effectively improved.

The invention provides a Maxout activation function operation acceleration device based on a storage and calculation integrated device and a method of the Maxout activation function operation acceleration device. The device comprises a data splitting module used for splitting data into an integer part and a decimal part; the data interception module is used for splitting data of an integer part and a decimal part into fixed bit lengths in a binary form, transmitting the data to the storage and calculation integrated device array, generating corresponding shift control signals and transmitting the shift control signals to the accumulator module; the comparator module is used for comparing output results of the array to obtain a maximum value and outputting the maximum value; the accumulator module is used for performing shift accumulation on an output result of the comparator module and outputting an accumulated result according to a shift control signal; and the weight queue storage module is used for storing weights corresponding to neural network workloads and updating the weights stored and calculated in the integrated device array after one layer of calculation is completed. The method has the advantages of low power consumption, high energy efficiency, adjustable precision and low calculation delay.

The invention discloses an intelligent glasses double-target speech enhancement method and system based on a multi-microphone array, terminal equipment and a medium, and relates to the technical field of speech enhancement. The method is applied to the intelligent glasses, and comprises the following steps: collecting noise-containing multichannel original time domain voice signals through a multi-microphone array of the intelligent glasses, and performing time-frequency domain conversion to obtain a frequency spectrum set; frequency domain complex spectrum amplitude-phase characteristics of each channel are extracted, a local time sequence context is introduced, and multi-channel space acoustic characteristics of amplitude-phase decoupling are constructed; and regulating and controlling the double-tone-area complex spectrum through pre-trained deep neural network modeling to obtain a target complex spectrum, and performing inverse time-frequency domain conversion and reconstruction to obtain an enhanced independent wearer voice signal and a preset direction target voice signal. According to the method, double-target voice parallel enhancement and non-target interference suppression are realized, the power consumption is low, the real-time performance is high, the method adapts to the harsh performance constraint of the end side of the intelligent glasses, and the robustness of voice enhancement in a complex noise scene is excellent.

The application discloses a same-row relationship detection method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining target video data in a target place; determining a personnel pixel area in each image frame in the target video data, and determining a target height and / or a target body shape ratio of the personnel corresponding to the personnel pixel area; determining an anonymous identity of the personnel in the image frame based on a dynamic anonymous personnel name list corresponding to the target place and the target height and / or the target body shape ratio; and judging whether a first search personnel and a second search personnel are in a same-row relationship according to the anonymous identity information of the personnel in each image frame in the target video data. The scheme can efficiently and with low algorithmic consumption detect the anonymous identity of the personnel in each image frame in the video data, so as to accurately and quickly find out whether the two search personnel are in a same-row relationship according to the anonymous identity of the personnel.

The invention relates to the technical field of Internet of Things communication, discloses an Internet of Things multi-Android client state synchronization method, and solves the problems of poor state consistency, high back-end dependence, large delay and insufficient compatibility in the prior art. According to the scheme, the method comprises the steps that after a client completes initialization, MQTT connection with a rear end is established, an MQTT state message pushed by the rear end is monitored, after the message is received, equipment state information is obtained through analysis, local state cache and a UI interface are updated, meanwhile, the client serves as a distribution node, a UDP synchronization message containing the equipment state and the IP of the distribution node is established, and the UDP synchronization message is sent to the client. The UDP synchronization message is broadcasted to the local area network, other clients in the local area network receive the UDP synchronization message and then verify whether the distribution node IP is the client in the local area network, if the verification is passed, the message is analyzed, the state updating timestamp and the local cache timestamp are compared, and the local state cache and the UI interface are updated only when the state updating timestamp is larger.

Embodiments of the present application provide a check matrix generation method, device and electronic equipment. The check matrix generation method comprises: generating a plurality of reference check matrices based on target configuration parameters and error correction performance requirements; generating a plurality of mapping tables corresponding to different parameter combinations based on the target configuration parameters, constraint conditions of an LDPC encoder and constraint conditions of an LDPC decoder; adjusting the plurality of reference check matrices based on the plurality of mapping tables to obtain adjusted check matrices; and checking the adjusted check matrices, and taking an adjusted check matrix that passes the check as the target check matrix, which can shorten the check matrix generation time and reduce the computing power consumption.

The invention relates to the technical field of disaster early warning processing, in particular to a mountain torrent disaster early warning system based on deep learning and air-space-ground integrated perception. The system comprises the following modules: a space-air-ground sensing module, an edge calculation preprocessing module, a deep learning prediction module and an intelligent decision and early warning module. Wherein the air-space-ground sensing module is used for collecting multi-source data; wherein the multi-source data comprises earth surface deformation data obtained through a spaceborne synthetic aperture radar, high-precision topographic data obtained through an unmanned aerial vehicle laser radar, hydrological sensing data obtained through a ground internet-of-things distributed sensor, and meteorological numerical forecasting data obtained through a meteorological data interface. Through air-space-ground integrated perception, physical constraint deep learning and edge intelligence technologies, mountain torrent multi-source data collaborative acquisition and accurate mountain torrent disaster prediction are realized, so that the comprehensiveness, reliability and timeliness of mountain torrent disaster early warning are improved.

The invention provides a matrix calculation method and device and related equipment, and relates to the field of digital circuits. The matrix calculation method is applied to a processing unit. The matrix calculation method comprises the following steps: acquiring a target bit width region of a target element in matrix calculation; the target element is an element which needs to be subjected to bit value adjustment in the matrix calculation; the target bit width region is a bit width region which can still enable the target element to meet the calculation precision requirement corresponding to the target element after the value on each bit is modified; adjusting the value of a target bit width area of the target element to be 0; and performing matrix calculation based on the adjusted target element. Through the method, the power consumption of matrix calculation can be reduced on the premise that the matrix calculation meets the precision requirement.

The present invention discloses a method for training an AI system based on big data, including: obtaining basic data for training an AI system, establishing a basic database based on the basic data; receiving an AI system to be trained sent by an AI developer; obtaining attribute information of the AI ​​system to be trained; obtaining target data based on the attribute information in the basic database; and performing a training task on the AI ​​system to be trained based on the target data. This solution can effectively ensure the universality of training data in the AI ​​system training industry, reduce the waste of computing power caused by the non-universality of training data, and can also form a universal third-party training system that is independent of AI system developers and users, further lowering the threshold of the AI ​​system training industry, and is conducive to the advancement of AI system technology in today's society.

The invention relates to a living body detection method and device, an electronic device and a storage medium, and is applied to the field of target recognition, and the method comprises the steps: inputting a to-be-detected image collected by monocular equipment into a target lightweight network structure in a living body detection network, and obtaining a face detection frame and a living body label corresponding to the face detection frame; and the target lightweight network structure is used for acquiring target global features, including dense semantic information in the to-be-detected image, in the to-be-detected image. Cutting a human face detection frame corresponding to the target living body label containing the living body content to obtain a cut human face, and inputting the cut human face into the target texture discrimination network to obtain a living body detection result; and the target texture judgment network is used for carrying out classification judgment on the clipped face on a texture feature level. According to the method and the device, the problem that hardware cost saving and accurate face detection effect keeping cannot be taken into consideration under the condition that only one-way color image input exists in the prior art is solved.

This invention belongs to the field of power monitoring technology and discloses a brightness balancing and regional smoothing fusion algorithm for power monitoring images. The method includes: image denoising, using mean filtering and median filtering to smooth the image; image registration, performing optimal matching on two or more images obtained from different sensors, at different times, or from different angles; and image region fusion, smoothly fusing images with different angles and brightness contrasts. This method is applicable to panoramic video monitoring systems. Designed specifically for the complex lighting and high electromagnetic interference characteristics of power operation sites, this invention effectively eliminates on-site noise through adaptive preprocessing, and achieves a balance between global brightness uniformity and local detail preservation through layered brightness balancing, solving the problems of incomplete brightness balancing and detail loss in power scenarios encountered by general algorithms.

The invention discloses a reconfigurable logic unit based on a ferroelectric transistor, a control method and a reconfigurable logic system, and belongs to the technical field of integrated circuit design. The reconfigurable logic unit comprises a P-type MOSFET (Metal-Oxide-Semiconductor Field Effect Transistor), a first N-type FeFET, a second N-type FeFET and a third N-type FeFET; wherein the first N-type FeFET and the second N-type FeFET are connected in parallel, drain electrodes of the first N-type FeFET and the second N-type FeFET are connected with a drain electrode of the P-type MOSFET, and source electrodes of the first N-type FeFET and the second N-type FeFET are connected with a drain electrode of the third N-type FeFET; the source electrode of the third N-type FeFET is grounded; according to the reconfigurable logic unit, the grid electrode of the N-FeFET serves as an input terminal, the polarization characteristic of a ferroelectric device is fully utilized, high-state threshold voltage and low-state threshold voltage can be separated under the control of polarization voltage, logic reconfiguration is achieved, and the reconfigurable logic unit has the advantages of being small in layout area, low in calculation power consumption, high in reconfiguration capacity, compatible with a CMOS, easy to process integration and the like.

The invention relates to the technical field of vehicle control, and discloses a vehicle starting control method and device, a vehicle and a medium, and the method comprises the following steps: in response to starting operation of the vehicle in an automatic parking state, adjusting parking release control parameters according to current road surface information and vehicle parameters; and releasing the automatic parking state according to the adjusted parking release control parameter. The parking release control parameters are dynamically adjusted according to real-time road surface information and vehicle parameters through a dynamic and self-adaptive parameter adjusting mechanism, the limitation of traditional fixed calibration is eliminated, different road conditions and vehicle states can be adapted, the vehicle starting quality in the automatic parking state is remarkably improved, and the vehicle starting speed is increased. And fixed parameters under the automatic parking function can be continuously optimized, so that the method adapts to individual manufacturing differences of different vehicles, starting abnormities such as slope sliding, dragging and vehicle channeling of the vehicles can be effectively avoided, the safety and smoothness of vehicle starting are improved, and meanwhile the user driving experience and the vehicle intelligent control level are greatly improved.

The application relates to the technical field of laser precision machining, in particular to an optical coherence tomography real-time feedback control system based on neuromorphic computing, which comprises an optical coherence tomography device and a processing module; the optical coherence tomography device comprises a collecting module and an executing module; the collecting module is used for collecting an analog voltage signal corresponding to an original optical interference signal of a biological tissue; the processing module is used for converting the analog voltage signal into a digital signal stream, inputting the digital signal stream into a neural network model, outputting a target depth position index from the neural network model, and generating a control signal according to the target depth position index; and the executing module is used for controlling the energy output of a surgical laser guided by the optical coherence tomography according to the control signal, simulating a reflex arc mechanism of a biological nervous system, constructing a bionic optical reflex arc architecture, and being low in calculation delay and small in computing power consumption. Therefore, the problems that related technologies cannot match microsecond laser pulses and cannot realize precise control are solved.

The invention relates to a dialogue participant identification method and system based on intelligent glasses. The intelligent glasses are provided with a bone conduction microphone and an air conduction microphone; the bone conduction microphone is arranged at a position in contact with the body of a wearer in a wearing state and is used for collecting bone conduction sound signals; the air conduction microphone collects an air conduction sound signal. According to the dialogue participant identification method based on the intelligent glasses, the bone conduction sound signal is used as the wearer sound signal, the wearer sound signal is subtracted from the air conduction sound signal, and the participant sound signal is obtained. The same dialogue participant recognition effect can be achieved without operating a neural network voiceprint recognition model with many parameters and a complex structure, and the method has the advantages of being small in computing power consumption, short in recognition time, good in real-time performance, high in recognition accuracy and high in cost performance.

The invention discloses an aerial writing method and device, a storage medium and electronic equipment, and is applied to an aerial writing pen, and the method comprises the following steps: determining a spatial three-dimensional coordinate of the aerial writing pen based on a wireless positioning communication signal between the aerial writing pen and corresponding display equipment; calculating a screen two-dimensional coordinate of the aerial writing pen based on the spatial three-dimensional coordinate; and sending the screen two-dimensional coordinates to a display device, so that the display device displays a cursor corresponding to the aerial writing pen based on the screen two-dimensional coordinates. In the movement process of the aerial writing pen, the movement position change between the aerial writing pen and the display device can be sensed based on the wireless positioning communication signal, the space three-dimensional coordinate of the aerial writing pen is obtained, then the screen two-dimensional coordinate of a display cursor of the aerial writing pen in the display device is determined, the calculation power consumption needed for processing the wireless positioning communication signal is low, and the calculation efficiency is high. Accurate aerial handwriting positioning can be completed with less computing power.

The application discloses a kind of in-service energy storage battery performance evaluation method, belong to the technical field of energy storage battery safety monitoring.Evaluation method is by analyzing the overall data characteristics of energy storage battery cluster, constructs distributed multi-scale deep learning joint statistical regression algorithm, forms experience model;Wavelet analysis fusion addition counter is constructed, according to the power unit time cumulative principle, the cumulative power difference is calculated to determine the relative output difference of single compared with the remaining battery in the same group.Simultaneously, because the current of series battery is equal, only voltage difference exists, so current value data is not used, and voltage is used for monitoring.The application fuses deep learning experience model, wavelet analysis method, and combines the difference of vehicle-mounted computing power and traditional CPU, reduces the computing power consumption, vehicle-mounted energy consumption and heat stress, improves the safety performance of energy storage battery pack, prolongs the overall calendar life of battery, improves the battery resource utilization efficiency, and reduces the life cycle cost of battery.

The invention provides a streaming media code adjustment method and device, electronic equipment and a storage medium, and relates to the technical field of intelligent cabins. The method comprises the following steps: when a cloud end pushes flow to at least one screen of a client, judging whether computing power resources of the client are sufficient or not; if the computing power resources are sufficient, configuring streaming media pushed by the cloud to each screen of the client to apply a first coding level; if the computing power resources are insufficient, according to the computing power priorities of all the screens of the client, the coding levels of streaming media pushed by the cloud to all the screens of the client are sequentially reduced, and the computing power resources needed by the streaming media are in positive correlation with the coding levels of the streaming media. According to the invention, the coding level of the streaming media pushed to each screen of the client by the cloud is reduced in sequence, so that the computing power consumption of the client can be reduced in sequence, the display requirement and the required performance of the client are met under the condition of limited computing power, and the user experience is ensured.

The invention discloses a black vehicle-oriented real-time point cloud fusion detection method, system and device, and a medium. The method comprises the following steps: step 1, obtaining point cloud data through a laser radar; 2, scanning the point cloud data, marking the continuous point cloud blocks with the reflection intensity of 0 as unlicensed car identifiers, and storing the unlicensed car identifiers in point cloud attribute positions; step 3, performing dual-thread compensation, performing clustering operation on the point cloud data set with the unlicensed car identifier by using DBSCAN, and performing imaging processing on a clustering operation result by using Point Pill to obtain a three-dimensional image model; forcibly outputting an unlicensed vehicle detection frame image according to an image processing result or improving the confidence coefficient as an unlicensed vehicle compensation confidence coefficient, and preferentially adopting a model result; when the two conditions do not belong to the above two conditions, performing weighted fusion on the operation result; and step 4, fusing operation results in the step 3 into a complete three-dimensional frame image. Therefore, the problem of missing detection is solved, and the detection rate and detection efficiency are improved.