200results about How to "Quick snap" patented technology

The invention discloses an intelligent fault diagnosis method based on a multi-input neural network. The method comprises the steps of data acquisition, preprocessing, feature extraction, model training and fault judgment. A variety of sensor signals are collected from the device, de-noised and normalized to generate high quality features. Time domain and frequency domain features are extracted in combination with methods such as wavelet transform, deep learning is performed on feature data by using a neural network, and a potential fault mode is identified. Through the trained model, the system can detect and classify faults in real time, output fault levels and diagnosis reports, and provide support for equipment maintenance. The method is suitable for a large-scale complex environment, has high diagnosis precision and quick response capability, effectively improves the management efficiency, and reduces the non-planned downtime.

The invention discloses a real-time three-dimensional fusion reconstruction method for a ship comprehensive test scene, and the method comprises the steps: carrying out the preprocessing of a target object image data flow, converting an image coordinate system into a camera coordinate system, obtaining a sparse point cloud model, and constructing a voxel three-dimensional grid; the method comprises the following steps: constructing a scale space by using a Gaussian difference pyramid to detect key points of an image under different scales, calculating a gradient direction histogram around the key points and taking the gradient direction histogram as a descriptor, matching the descriptor by using an Euclidean distance, and estimating relative pose transformation between the images to complete a local matching process. Projecting and mapping the local frame block to a reference frame to form a key frame, and entering a global matching process; and extracting a surface grid of the three-dimensional model from the dense body grid by adopting a Marching Cubes algorithm, traversing each cube unit in the surface grid, and obtaining distance information from a vertex to a nearest surface based on the cube units, thereby generating a final three-dimensional model of the test target object.

The invention relates to the technical field of medical data analysis, in particular to a lung cancer dynamic evaluation method and system based on immune monitoring, and the method comprises the following steps: collecting immune index data obtained in the immune monitoring process of a patient, analyzing the mean value change between adjacent segments of an immune index data time sequence, marking an abnormal fluctuation region, and calculating the immune index data; and generating abnormal fluctuation section data. According to the method, the immune monitoring data of the patient are divided into the time sequences, the abnormal fluctuation sections and the abnormal change points are accurately extracted through the mean value change analysis and sliding window method, the remarkable fluctuation of the immune system can be captured more rapidly and accurately, and therefore the timeliness of dynamic monitoring is improved. And further in combination with statistical feature analysis and similarity matrix calculation, differentiated grouping of immune indexes among patients is refined, and personalized evaluation for patient groups with different immune states is facilitated.

The invention relates to the technical field of enterprise cloud management, and discloses an ERP enterprise cloud management system, which comprises a system main body deployed based on a cloud computing architecture, the system main body is integrated with a financial management module, an inventory control module, a purchase management module, a sales management module, a production plan module, a human resource management module and a customer relationship management module, and data interaction interfaces and cooperative work mechanisms are arranged among the modules; data are stored by the cloud storage unit, and a data processing task is executed by the cloud operation unit. The ERP enterprise cloud management system breaks data barriers among departments of an enterprise, integrates core modules of financial management, inventory control, purchase management, sales management, production planning, human resource management, customer relationship management and the like, ensures the accuracy and consistency of data, and realizes the cloud management of the enterprise through a data sharing function. All the modules can obtain needed data and information in real time, and seamless transmission and cooperative work of the information are achieved.

The invention belongs to the technical field of new energy equipment anomaly analysis, and relates to a variable pitch component anomaly detection method and system and a readable medium, and the method comprises the following steps: calculating the actual pitch angle standard deviation of a blade, and recognizing whether the variable pitch angle is abnormal or not through the actual pitch angle standard deviation; if the variable pitch angle is not abnormal, the standard deviation of the collected data of all the measuring points is obtained, and whether measurement abnormity exists in all the components or not is judged one by one according to the standard deviation of the collected data; if the measurement abnormity does not exist in each part, calculating the slope of the standard deviation of the collected data, so as to judge whether the standard deviation of the collected data has a rising trend or not and judge whether a gradual deterioration risk exists or not; and if the deterioration risk does not exist, calculating a correlation matrix among the measuring points under various working conditions through the established working condition model, and judging whether the variable pitch component is abnormal under various working conditions according to the correlation matrix. According to the method, more complete and careful abnormal detection of the variable-pitch component can be realized.

The invention provides a data retrieval system and method based on a large language model and a readable storage medium. The system comprises a data analysis module, a multi-mode integration module, a semantic retrieval module and a post-processing module which are arranged in sequence. The data analysis module analyzes the original engineering data to generate structured data; the multi-modal integration module maps the structured data to the same semantic vector space to construct a vector database, and establishes a cross-modal association index to construct a multi-modal engineering knowledge graph; the semantic retrieval module is based on a large language model and a retrieval enhancement technology, responds to a natural language query of a user, retrieves related information from a vector database and a multi-modal engineering knowledge graph and generates answers; the post-processing module performs abstract generation on the natural language answers based on a large language model and generates associated recommendation content. According to the method, the query accuracy, the retrieval efficiency and the user query interaction experience can be remarkably improved; and the scheme reuse rate can be improved based on the recommended content.

The invention provides a real-time prediction method and device for gas storage surrounding rock fracture, equipment and a medium. Comprising the following steps: acquiring an acoustic emission signal and a stress-strain parameter generated in a fracture process of a specified type of rock in surrounding rock of the gas storage; screening out target acoustic emission signals meeting training conditions based on acoustic emission characteristic parameters corresponding to the acoustic emission signals; determining a rock state stage label of the target acoustic emission signal based on the stress-strain curve of the stress-strain parameter corresponding to the target acoustic emission signal; based on the acoustic emission characteristic parameters of the target acoustic emission signal and the rock state stage label of the target acoustic emission signal, constructing a sample data set; training based on the sample data set to obtain a rock fracture prediction model; the rock fracture prediction model is adopted to process to-be-detected acoustic emission signals, collected in real time, of the target rock to obtain a rock state stage, and the to-be-detected acoustic emission signals are signals, collected by the piezoelectric ceramic sensor, of the target rock around the reinforced structure of the gas storage.

The invention relates to the technical field of slump detection of mixed building materials, in particular to a device and a method for detecting the slump of a mixed building material based on machine vision. The method comprises the following steps: marking collected video data to form a model training set and a test set; a video image frame is converted into a gray level image, an SURF algorithm is used and combined with an ORB algorithm to extract a feature value point set of the final video image frame, a FlannBasedMatcher algorithm is used to carry out feature value matching on adjacent frames, matching points are screened, relative displacement of the matching points is calculated, and abnormal values in the relative displacement of the matching points are eliminated; constructing a slump detection model by utilizing a polynomial fitting technology and combining with Elastic Net regularization; and calculating the slump in real time based on the new video data in combination with a slump detection model. The slump detection model is constructed by combining machine vision feature point extraction and an Elastic Net regularization model, and the slump of the mixed material can be accurately captured in real time.

The invention discloses a visual characterization method for carbon emission of a sewage treatment plant, belongs to the field of carbon emission monitoring, and solves the problem that the fluctuation and trend of carbon emission data cannot be analyzed by an existing method. The method comprises the following steps: checking carbon emission intensity; preprocessing the carbon emission intensity data; defining a color mapping function according to the numerical range of the carbon emission, and mapping the numerical value to a color space; based on the color mapping function, creating a color bar for representing a corresponding relationship between the data value and the color; according to the color mapping function and the color bar, converting the carbon emission data at each time point into a corresponding color, and drawing the color on the heat map; and optimizing and outputting the image. According to the method, the calorific value map technology is introduced, so that the visualization effect of carbon emission intensity data is remarkably improved, the change of the carbon emission intensity in a time sequence is visually displayed, complex data becomes clear at a glance, and the peak value and the fluctuation trend of carbon emission can be quickly captured.

The invention relates to the technical field of ship automatic control and information safety, solves the technical problem that a traditional method cannot effectively monitor in real time and quickly give an early warning to prevent various accidents when an inland ship is in a complex and changeable operating environment, and particularly relates to an abnormal state detection method and system for a shore-based driving and controlling inland ship group. The method comprises four parts: identity authentication, node anomaly detection, dynamic trust evaluation and node adaptive control. According to the method, only reliable entities can participate in communication through a strict verification process and continuous trust evaluation, so that communication safety is enhanced, comprehensive and real-time monitoring of inland ship groups is realized, abnormal conditions such as physical faults or network attacks can be found in time, potential risks can be warned in advance, and the inland ship groups can be monitored in real time. Therefore, the safety and the reliability of the whole navigation are improved.

The invention provides a personalized recommendation method and system for hot events, and relates to the technical field of big data analysis, and the method comprises the steps: collecting multi-source real-time data; preprocessing the multi-source real-time data to obtain multi-source real-time preprocessed data; performing weighted fusion on the multi-source real-time preprocessing data to obtain a weighted fusion score; generating a hotspot event based on the weighted fusion score; and retrieving the hot events to generate personalized recommendation contents meeting user requirements. According to the invention, real-time and intelligent generation of the hotspot event is effectively realized, personalized hotspot event recommendation is provided for the user, and the user experience is improved.

The invention relates to the technical field of fault diagnosis, and provides a nuclear power station equipment fault diagnosis method and system, which can reflect the state of each piece of equipment and the health condition of the whole system in real time by collecting equipment operation data in real time and dynamically constructing and updating an equipment relation graph by using physical connection and function dependency relationship among the equipment. Besides, probability calculation of equipment fault influence is introduced in the method, and the possible influence of the fault equipment on other parts of the system can be accurately analyzed based on the total score of the equipment and the dependency relationship between the equipment, so that quicker and more effective response is made; the problem that an existing method is difficult to accurately analyze the complex dependency relationship and the interaction influence between devices is solved.

The invention relates to the technical field of computer machine vision, and discloses a plant growth condition monitoring method, system and device and a storage medium, and the method comprises the steps: obtaining a target plant image, inputting the preprocessed target plant image into an improved YOLOv9 model (comprising a backbone network, a feature fusion network and a prediction classification network), and obtaining a target plant image; obtaining a target plant detection image; performing feature extraction on the preprocessed target plant image through a backbone network to obtain plant multi-scale features; performing feature conversion, branch processing and feature fusion on the plant multi-scale features through a feature fusion network to obtain plant fusion features; performing classification processing on the plant fusion features through a prediction classification network to obtain a target plant detection image; and inputting the target plant detection image into an image semantic segmentation model (including an image super-resolution network and an image segmentation network) to obtain a plant binary image. The method effectively improves the accuracy of plant growth condition monitoring.

The invention relates to the field of stacking and sorting, in particular to a multi-specification product mixed unstacking and intelligent sorting device based on three-dimensional visual guidance, which comprises a sorting robot, the sorting robot is provided with a mounting frame arranged at the driving end of the sorting robot and an adsorption assembly arranged on the mounting frame, and the periphery of the mounting frame is provided with an identification camera; the anti-interference mechanism comprises a curtain blocking assembly capable of blocking stacks and a distance adjusting assembly used for adjusting the distance between the curtain blocking assembly and the stacks, the curtain blocking assembly comprises a fixed part, a movable part and a curtain fixedly connected between the fixed part and the movable part, and the distance adjusting assembly is used for adjusting the distance between the curtain blocking assembly and the stacks. The sucked stacks are surrounded by the four curtains, so that a space for shielding the stacks in the space is formed, external light and other factors possibly interfering the identification process are effectively isolated, then the stacks are shot through the identification cameras on the periphery of the stacks, and the specific specification of the stacks is accurately judged.

The invention discloses a Java-based intelligent document generation method, and belongs to the technical field of intelligent generation, and the method comprises the steps: S1, constructing a deep learning model: constructing a natural language processing model based on large-scale legal text data set training by using a deep learning framework integrated by a Java platform, the natural language processing model is used for understanding complex legal concepts, terms and logic relations, in the method, deep law understanding ability is achieved, deep understanding of legal texts is achieved through application of the deep learning model, a shallow processing mode of traditional template filling is exceeded, efficiency is improved, the legal document preparation time is greatly shortened, and the method is suitable for popularization and application. The workload of legal professionals is reduced, the accuracy is enhanced, document errors and omissions are reduced, the specialty and compliance of legal documents are ensured, standardization is promoted, the standardized process of legal document writing is pushed, the consistency and quality of legal services are improved, personalized services are achieved, and the special requirements of specific clients or cases are met.

The invention discloses a test device for simulating migration and transformation of pollutants in underground water. The test device comprises a sand box, a water supply system and a drainage system, the sandbox is internally divided into a water inlet well, a water outlet well and a sand tank; the water inlet well adjusts water head height and is connected with the water supply system; the water supply system, the water inlet well and a passage therebetween are sealed; the sand tank simulates an aquifer soil body and is filled with a soil sample or quartz sand; the water outlet well is connected with a drainage system; the water inlet well and the water outlet well are separated from the sand tank through partition plates, a plurality of through holes are formed in the partition plates, and filter cloth is arranged on the surfaces of the partition plates facing the sand tank. A plurality of pore water sampling holes are formed in the front upper part of the sand tank; a plurality of underground water sampling holes are formed in the front lower part of the sand tank; a plurality of pressure measuring holes connected with a pressure measuring pipe or inserted into a soil moisture tensiometer are formed behind the sand tank; water of the water supply system enters the water inlet well and then permeates into the sand tank filler, water in the filler permeates into the water outlet well, and water of the water outlet well is discharged from the drainage system. Compared with a field test, the material and time cost is reduced.

The invention discloses a carbon fiber crushing device and a crushing method, and belongs to the technical field of carbon fiber crushing devices.The carbon fiber crushing device comprises a crushing box and two crushing rollers rotationally installed in the crushing box, material boxes are welded to the two sides of the crushing box correspondingly, and two sliding plates are arranged in the material boxes in a sliding mode; a plurality of pressing plates are arranged between the two sliding plates, and every two adjacent pressing plates are arranged in a V shape. According to the carbon fiber braided fabric crushing device, the arranged compression assembly is matched with the cover plate, the carbon fiber braided fabric can be compressed, in the process that the sliding plate is moved out of the crushing box, every two adjacent pressing plates can be separated, so that the carbon fiber braided fabric can fall to the crushing rollers, and the carbon fiber braided fabric is compressed into a triangular prism shape, so that the carbon fiber braided fabric can be quickly captured by the two crushing rollers; when the sliding plate is located between the cover plate and the material box, dust can be prevented from overflowing in the crushing process.

The invention discloses a reservoir dam flood discharge affected river reach safety early warning system and an early warning method thereof, and belongs to the technical field of dam flood discharge safety early warning of water conservancy and hydropower engineering. The system comprises an early-warning river reach and a plurality of unmanned aerial vehicle monitoring stations, a plurality of fixed targets are arranged on bank slopes on the two sides of the early-warning river reach, the fixed targets on the bank slopes on the two sides of the early-warning river reach are alternately arranged in a staggered mode, the unmanned aerial vehicle monitoring stations are arranged on the bank slopes of the early-warning river reach, and the unmanned aerial vehicle monitoring stations are provided with unmanned aerial vehicles. A high-definition camera, an RTK measuring instrument, a wireless transceiver module and a controller are carried on the unmanned aerial vehicle. The unmanned aerial vehicle can quickly capture a fixed target, and can quickly and accurately position a dangerous case position in combination with a high-precision RTK measuring instrument carried on the unmanned aerial vehicle, so that the position of a person involved in danger can be confirmed in time when the dangerous case occurs, the response timeliness of emergency rescue measures is improved, effective rescue measures are adopted in a targeted manner, and the emergency rescue efficiency is improved. And the risk of personnel casualties is reduced.

The invention discloses a pre-hospital registration method for intelligent management planning, and relates to the field of hospitalization management, and the method comprises the following steps: 1, obtaining the access authority of a hospital electronic health recording system, and reading the basic personal information and past medical history of a target patient; 2, recording current personal symptoms and personal signs of the target patient in real time, and obtaining a treatment scheme made by a current doctor for the target patient; 3, constructing an analysis model through a deep learning algorithm, obtaining historical medical record data through a hospital electronic health recording system, and extracting tag features of personal information, medical history, symptoms, signs, treatment schemes, hospitalization time and treatment effects of a plurality of patients; through multi-dimensional analysis of basic information, medical history, symptoms and treatment schemes of a patient, the analysis model can generate a more accurate illness condition severity coefficient, and through combination of the illness condition severity coefficient and real-time information of a nursing bed, an appropriate bed and a nursing level are automatically matched.

The invention discloses a calcium-zinc stabilizer, a preparation method and application. Relates to the technical field of calcium-zinc stabilizers. The calcium-zinc stabilizer provided by the invention comprises the following components in parts by weight: 25-40 parts of calcium salt; 15 to 25 parts of zinc salt; 2 to 5 parts of attapulgite; 0.1 to 0.4 part of graphene oxide; and 5-8 parts of an antioxidant. According to the preparation method of the calcium-zinc stabilizer, the attapulgite and the graphene oxide are mixed, and a mixture of the attapulgite and the graphene oxide is obtained through electrochemical deposition; mixing calcium salt, zinc salt and the mixture, adding water, and stirring to obtain a mixed solution; and adding an antioxidant into the mixed solution, uniformly mixing, drying by distillation to obtain a calcium-zinc stabilizer solid, and grinding to obtain the calcium-zinc stabilizer. The calcium-zinc stabilizer is applied to the field of PVC preparation. The attapulgite and the graphene oxide are added, channels are formed in PVC, calcium and zinc ions and the like rapidly move and are rapidly distributed on the surface and the interior of the attapulgite through a self porous structure, free radicals are rapidly captured, and the heating stability of all parts in the PVC is improved.

The embodiment of the invention relates to the field of Java application performance analysis, and discloses a thread stack performance analysis task processing method and system. The method comprises the steps that according to historical performance analysis data, the occurrence frequency of thread stack entry endpoints is determined; adaptively adjusting an initial weight value in a current thread stack performance analysis task according to the frequency to obtain a target weight value; wherein the initial weight value is used for representing a uniform weight value set in an initialization stage of the thread stack performance analysis task; according to the target weight value, adaptively adjusting an initial snapshot sampling rate to obtain a target snapshot sampling rate, and processing a thread stack performance analysis task according to the target snapshot sampling rate; wherein the initial snapshot sampling rate is used for representing a unified sampling rate set in an initialization stage of the thread stack performance analysis task. The technical problems that in the prior art, the burden of data processing is increased, and computing resources and storage space are wasted can be at least solved.

The application provides a kind of operation and maintenance scene visualization report generation method and system based on natural language question and answer, belongs to intelligent operation and maintenance technical field, method includes: receiving the natural language query input by user;Through the pre-training of large language model and the preset operation and maintenance terminology dictionary, the natural language query is parsed and entity is extracted, and the entity-field association table containing demand type is generated;According to demand type and semantic feature, the query type is judged to be data query or visual query;Based on the judgment result, generate structured query language sentence or visual instruction;Query is executed to the interface business database to obtain raw data;Raw data is processed and analyzed to obtain analysis result data;Call data feature adaptation algorithm to match chart type, generate and output visual report.The application realizes the full-link automation from natural language input to visual report output, reduces the operation and maintenance data interaction threshold, improves the operation and maintenance data processing efficiency and accuracy.

The invention provides a machine vision-based moving target detection method and apparatus, and an electronic device. The method comprises the steps of obtaining an original image by using an installed camera; based on a pre-trained edge detection model, extracting edge information of the original image under different scales to obtain a multi-scale edge image; and based on a pre-trained shape model, obtaining a moving target detection result according to the multi-scale edge image. According to the method, the target object is positioned through the shape model, the dependence on the appearance of the target can be reduced, the robustness of target positioning can be improved, meanwhile, the accuracy and stability of edge detection can be effectively improved by improving a traditional edge detection algorithm and extracting the edge information of the original image under different scales, and the accuracy and stability of target positioning can be improved. Therefore, the position of the target can be captured more quickly and accurately, and the category of the target can be predicted.

This utility model relates to the field of manifold technology and discloses a gas manifold with gas leak detection function. The novel solution includes a gas manifold, with a mounting cover fitted around the gas delivery pipe. An externally threaded connecting ring is integrally formed at the bottom of the mounting cover, and a sealing sleeve is screwed onto the outside of the externally threaded connecting ring. A magnetic ring is embedded at the bottom of the sealing sleeve. This utility model, by setting the mounting cover and sealing sleeve at the connection between the gas delivery pipe joint of the gas manifold and the mouth of the gas storage cylinder, completely covers the joint where it meets the mouth of the gas storage cylinder, forming a closed detection space. This confines the leaked gas to a small area. Combined with a concentration detection sensor on the mounting cover, it enables rapid detection and source location of leaks. Simultaneously, an independent audible and visual alarm is installed on the outside of the mounting cover to ensure that operators are alerted immediately to pinpoint the fault when a leak occurs.

The invention belongs to the field of detection of canine heart injury, and particularly relates to a kit for detecting canine heart injury and application. The kit comprises a probe, a conjugate or the kit can realize cyclic amplification under the assistance of DSN, and target miRNA can be quickly captured. The probe, the conjugate and the kit provided by the invention can be used for detecting a dog serum sample and judging a dog heart injury condition according to a serum detection result, and the accuracy is higher than that of a conventional method.

The invention relates to a signal control circuit based on a transverse tandem photoelectric detection unit structure and a control method, and belongs to the field of photoelectric detection and control. The circuit comprises a transverse tandem detection unit structure and a post-stage signal processing circuit, wherein the post-stage signal processing circuit comprises a plurality of resistor units, a bipolar transistor and an MOS (Metal Oxide Semiconductor) tube; according to the method, an optical signal is converted into an electric signal through a transverse tandem type detection unit structure, and charge release and voltage signal output processing are carried out through a post-stage signal processing circuit. When light enters the transverse tandem type detection unit structure, the output tube is switched on; when no photo-generated current exists, the output tube is switched off; and the output tube can output a voltage control signal meeting the amplitude requirement when being switched on or switched off so as to control the working state of a rear-end circuit. By combining the photoelectric detection unit and the signal processing circuit, monolithic integration and process fusion of the silicon-on-insulator process are realized, the parasitic effect between elements can be effectively reduced, and the interconnection scheme of discrete devices is upgraded.

The utility model discloses a parking device for butt joint of a pallet and a space station, which relates to the technical field of parking for butt joint of space stations and comprises a driving motor, the output end of the driving motor is fixedly connected with one end of a driving rotating shaft, and the middle of the outer circle surface of the driving rotating shaft is slidably connected with the middle of the lower surface of a lifting disc. The lower surface of the lifting disc is connected with a locking claw device, the locking claw device comprises a fixing frame, one end of the lower surface of the fixing frame is fixedly connected with one end of a first fixing rod, the upper end of one side of the first fixing rod is rotationally connected with one end of a first fixing pin, and the middle of the outer circle surface of the first fixing pin is rotationally connected with one end of a first sliding sleeve. According to the utility model, through the arrangement of the locking claw device, the driving motor drives the driving rotating shaft, so that the fixed claw can slide downwards according to the butt joint position to drive the locking claw device to fix and lock the parking contactor, and the situations of shaking and unstable butt joint can be effectively prevented when butt joint is carried out on a space station; and rapid capture can be realized.

The invention discloses a carbon fiber crushing device and a crushing method, belonging to the technical field of carbon fiber crushing devices, wherein the carbon fiber crushing device comprises a crushing box and two crushing rollers rotatably installed in the crushing box, material boxes are welded at both sides of the crushing box, two slide plates are slidably arranged in the material box, a plurality of pressure plates are arranged between the two slide plates, and a V-shaped arrangement is formed between two adjacent pressure plates; the carbon fiber braid can be compressed by cooperating with the cover plate through the arranged compression component and the cover plate, and the adjacent pressure plates can be separated in the process of the slide plate moving out of the crushing box, so that the carbon fiber braid can fall to the crushing roller, and the carbon fiber braid is compressed into a triangular prism shape, which is conducive to being quickly captured by the two crushing rollers. In addition, when the slide plate is between the cover plate and the material box, dust overflow during the crushing process can be prevented.

The invention belongs to the technical field of communication, and particularly relates to an OFDM (Orthogonal Frequency Division Multiplexing) end-to-end communication system with environment self-adaptive capability. According to the invention, the existing E2E communication system is combined with the CP-OFDM technology, so that the E2E system based on OFDM can effectively resist the multipath effect. In addition, the problem of environmental adaptability of the E2E system in practical application is solved by utilizing a meta-learning technology: firstly, pre-training of a noise simulation network is realized through an MAML technology, so that the noise simulation network can quickly adapt to a changing environment and capture environmental noise; thirdly, widely training the E2E receiving and transmitting network by using noise distribution captured by the channel simulation network, so that the E2E system has relatively high adaptability to channel fading and change of a received signal-to-noise ratio; finally, the OFDM-E2E communication system based on the environment adaptive training strategy can adapt to changes of three dimensions of channel noise, channel fading and a received signal-to-noise ratio, so that the performance close to the optimal performance is always kept in practical application.

The present application relates to a channel estimation optimization method, device, equipment and medium of a base station MIMO wireless communication system, which relates to the field of data processing, and the method includes: obtaining the channel preamble code in each communication connection request; based on the channel preamble code in the communication connection request, determining the pulse emission rate of multiple neurons corresponding to the channel preamble code in a preset channel estimation optimization model; in the neurons of the channel estimation optimization model, using Poisson distribution to perform Poisson sampling according to the pulse emission rate, randomly generating the time interval for the next pulse emission, and when the accumulated time interval reaches or exceeds the preset time interval threshold, recording the channel pulse code of the channel preamble code corresponding to the current time point; inputting the channel pulse code of the channel preamble code into the variational autoencoder in the channel estimation optimization model to reconstruct the channel preamble code. While improving the accuracy and efficiency of channel estimation, the present application also significantly reduces the energy consumption of the MIMO wireless communication system.