122results about How to "Reduce storage space" patented technology

The invention provides a multi-path embedded video coding method and system based on generative artificial intelligence. The multi-path embedded video coding method comprises the following steps: firstly, coding input cue words and demand parameters, and outputting cue word vectors and demand parameter code streams; then decoding the cue word vector and the demand parameter code stream, and outputting a feature vector and a feature map; extracting the feature vector, the feature map and the representation guide vector to obtain multi-embedded feature information; reconstructing a picture by using the obtained multi-embedded feature information through a generative AI and obtaining a video to be iterated; performing video quality evaluation on the video to be iterated; and generating a binary code stream for transmission through the current cue word vector, the demand parameter code stream and the multi-embedded feature information input. According to the method, important information characteristics are transmitted in a highly abstracted manner, and a high-definition video is reconstructed at a receiving end, so that the storage space required by video coding and the size of a transmission information stream are greatly reduced, and higher-definition transmission under a narrower bandwidth is realized or video coding under an extremely low bandwidth becomes possible.

The invention relates to the technical field of artificial intelligence, in particular to a model construction method for a GUI agent, a program product and a storage medium, and the model construction method for the GUI agent provided by the invention comprises the following steps: performing pre-training based on a general data set in the GUI field to obtain a base model; a user demonstration operation track is collected online, and a personalized task execution track is obtained; automatically labeling the task execution track to obtain a labeled track; performing task planning training on the base model based on the marked track to obtain a scene planner, and combining the base model and the scene planner to obtain a personalized model; based on the plug-in scene knowledge base, corresponding scene knowledge is retrieved as context supplement during personalized task action planning. The task execution track is automatically labeled, a large model is guided to actively label and understand thinking in human execution operation, the next behavior is further planned, and the adaptive capacity of a GUI intelligent agent to the personalized use environment of a user is greatly improved.

The invention relates to the technical field of video compression processing. The invention provides a monitoring video compression processing method and device, and the method comprises the steps: extracting the video features of a to-be-processed image in a monitoring video in a target scene region through a video feature extraction model, and obtaining the feature vectors of a plurality of video features in different image regions; according to the obtained feature vectors of the various video features in the different image areas, performing following classification on the to-be-processed image to obtain classification results: image area classification, scene type classification and importance level classification; determining an initial compression ratio of different image areas in the to-be-processed image according to the classification result; decomposing the to-be-processed image into sub-images of different scales by adopting an image decomposition method based on wavelet transform, and determining an optimal decomposition layer number according to image quality; and determining a final compression ratio and a coding mode of different image areas according to the scale and content characteristics of the sub-images of the image to be processed. According to the method, the compression ratio is greatly improved on the premise that the image quality is not obviously reduced.

The invention discloses a method for performing efficient lossless text compression based on a large language model, which relates to the technical field of text compression and comprises the following steps of: preprocessing to-be-compressed text data to obtain an initial token sequence; selecting and loading a pre-trained big language model, analyzing the current context information by using the big language model, obtaining a probability value of each token, forming a probability distribution list, and recording a ranking sequence number of the current token; the context window is moved backwards by one token, and context information is updated until all the tokens are processed; and recording the ranking sequence numbers of all the tokens in sequence to form a sequence number list so as to obtain a final text compression result. According to the method, the compression rate is remarkably increased, the requirements for storage space and transmission bandwidth are reduced, the high processing speed is kept, and the method is suitable for real-time compression and decompression of large-scale text data.

The invention discloses a voice attribute decomposition quantization system and method with a high compression ratio, and relates to the technical field of voice quantization, and the system comprises a generator composed of a feature extractor, an encoder, a decoupler and a decoder, and a discriminator. The feature extractor is used for extracting two-dimensional hidden features according to the audio signal to be quantized; the encoder is used for encoding the two-dimensional hidden features to generate hidden features; the decoupler is used for carrying out feature extraction on the hidden features to obtain a plurality of discrete tokens; the decoder is used for obtaining a reconstructed audio according to the plurality of discrete tokens; and the discriminator is used for assisting the generator in parameter updating. According to the method, through fine selection of the input features, enhancement of the expressive power of a decoder and introduction of various discriminators, generation of a voice quantization token with a high compression rate can be effectively guaranteed, on one hand, the data storage space and the LLM calculation amount are reduced, and on the other hand, modeling of a large model for an ultra-long audio can be promoted.

The invention provides an electric power visual knowledge distillation method and system, and the method comprises the steps: collecting electric power visual image data, and building an electric power visual data set; inputting the electric power visual data set into a first model for training to obtain a teacher model; constructing a second model, combining a detection framework of the teacher model with a backbone network of the second model, and constructing a student model; inputting the electric power visual data set into the teacher model and the student model at the same time to obtain output results, and calculating the difference degree between the output results of the teacher model and the student model by using a composite loss function; using a back propagation algorithm and an Adam optimizer to update parameters of the student model according to the difference degree; and enabling the difference degree to be smaller than a set value through a plurality of rounds of iteration, and outputting a student model. According to the method, on the premise that resource consumption is greatly reduced, the obtained student model has higher calculation speed, recognition precision close to that of a teacher model and smaller model size, and the electric power visual knowledge distillation system is suitable for various resource-constrained environments.

The invention relates to the technical field of Internet of Things, in particular to an Internet of Things intelligent linkage control method and an Internet of Things intelligent linkage control device, and the method comprises the steps: S1, obtaining the working state of home equipment and the real-time data of environment information, and transmitting the real-time data to a central control platform; s2, preprocessing the real-time data to obtain preprocessed data; s3, analyzing the preprocessed data, and generating a corresponding linkage control instruction; s4, encoding the linkage control instruction to generate encoded data, and performing lossless compression on the encoded data by adopting a ZSTD compression algorithm to obtain a compressed code; s5, searching an optimal sending communication network by adopting a flora optimization algorithm, and sending the compressed code to the home equipment; and S6, the home equipment decompresses the compressed code by using an LZMA compression algorithm, and executes the linkage control instruction. According to the invention, the ZSTD compression algorithm and the flora optimization algorithm are adopted, the delay of data transmission and processing is significantly reduced, the overall response speed of the smart home system is improved, and the user experience is better.

The invention relates to a content video-based hypercompression storage method. The method comprises the following steps of S1, acquiring video stream data of a camera through a network; s2, for the video stream data, performing corresponding decoding according to the code stream type to generate YUV data; s3, analyzing the YUV data, determining an ROI (Region of Interest), and extracting content feature values of a plurality of moving targets and position data of images where the moving targets are located as element data of the video; the ROI is logically separated from the non-ROI, secondary high-fidelity compression is carried out on the ROI, high compression is carried out on the non-ROI, and the high compression rate of the ROI under the high-fidelity condition is achieved; s4, combining the code streams of the ROI and the non-ROI into the same frame of code stream; s5, taking the ROI metadata of each region of interest as structured description data of the frame data; and S6, sending the coded code stream and the ROI metadata to a video storage engine for intelligent video storage.

The present invention relates to the technical field of video compression processing. A monitoring video compression processing method and device are provided, the method comprising: using a video feature extraction model to extract video features of an image to be processed in a monitoring video of a target scene area, and obtaining feature vectors of multiple video features in different image areas; performing the following classifications on the image to be processed according to the feature vectors of multiple video features in different image areas, and obtaining classification results: image area classification, scene type classification, and importance level classification; determining the initial compression ratio of different image areas in the image to be processed according to the classification results; using an image decomposition method based on wavelet transform to decompose the image to be processed into sub-images of different scales, and determining the optimal number of decomposition layers according to the image quality; determining the final compression ratio and encoding mode of different image areas according to the scale and content characteristics of the sub-images of the image to be processed. The present invention greatly improves the compression ratio without significantly reducing the image quality.

The invention relates to the technical field of computer storage, in particular to an address mapping method in storage equipment. Comprising the following steps: S1, setting a plurality of partitions with different mapping units in a storage device; s2, dividing the mapping table of the storage equipment according to the partitions in the step S1, wherein each partition corresponds to one sub-mapping table; s3, dividing the interior of the sub-mapping table into a plurality of mapping groups; s4, setting a mapping mode in the mapping group; and S5, when the mapping table is updated after the data is written, determining mapping table items partially overlapped with the written data LBA in the mapping table, and updating the first PBA and / or information representing PBA continuity in the mapping table items. Compared with the prior art, a plurality of partitions with different mapping units smaller than 4K can be arranged in the storage equipment and are used for improving the writing performance when small data blocks are accessed; and meanwhile, the mapping tables of different partitions can adopt a more compact mapping form according to the continuity of data writing, so that the storage space occupied by the whole mapping table is reduced, and the memory cost of the storage equipment is further reduced.

The invention discloses an atmospheric pollution three-dimensional monitoring system based on big data analysis. The atmospheric pollution three-dimensional monitoring system comprises a data acquisition module, a data preprocessing module, a sub-cluster generation module, an optimal cluster center number selection module, a cluster optimization module and an atmospheric pollution three-dimensional monitoring module. The invention belongs to the technical field of atmospheric pollution monitoring, and particularly relates to an atmospheric pollution three-dimensional monitoring system based on big data analysis, according to the scheme, setting of a threshold value and a judgment distance is introduced, and a point pair similarity matrix is selectively formed; the data storage space and the calculation amount are reduced; the convergence effect is accelerated; introducing a core data point; selecting an optimal clustering center based on the core data points and the contour coefficients; a search space is established based on clustering hyper-parameters, different movement strategies are made for individuals based on individual positions and optimized individual positions, and the convergence speed and search efficiency of the algorithm are improved; the three individuals with the lowest fitness values are reinitialized; and realizing atmospheric pollution three-dimensional monitoring based on clustering optimization.

The utility model discloses a scaffold for building construction, which belongs to the field of scaffolds and comprises a scaffold structure, two movable supports are symmetrically mounted on the scaffold structure, each movable support is composed of a sliding base plate, a fixed magnetic plate, a supporting plate, a connecting rod and a vertical rod, the fixed magnetic plate is fixedly mounted in the sliding base plate, and the supporting plate is fixedly mounted in the sliding base plate. The two supporting plates are symmetrically and fixedly installed at the outer end of the sliding base plate, the connecting rod is fixedly installed at the outer ends of the two supporting plates, the two vertical rods are fixedly installed at the two ends of the connecting rod respectively, and the movable support is arranged on the scaffold structure and the adjustable supporting structure is arranged on the movable support; when the scaffold structure is used, the movable support can be pulled outwards, then the adjustable supporting structure is rotated, the supporting contact plate makes contact with the ground, and therefore the using stability of the scaffold structure is improved, the scaffold structure is not prone to shaking and toppling over in the using process, and then potential safety hazards caused by using the scaffold structure can be reduced.

This invention provides a multifunctional bed frame that is easy to transport, relating to the field of furniture technology. It includes at least two multifunctional bed frames detachably mounted on a fixed frame. Each multifunctional bed frame includes a positioning frame and a lifting frame. The positioning frame is fixedly mounted on the upper side of the fixed frame, and the lifting frame is hinged to one side of the positioning frame and positioned on the upper side of the fixed frame. The lifting frame is raised or lowered by rotating upwards relative to one side of the positioning frame via a drive assembly. The advantage of this invention is that the modular structure, connecting the fixed frame to multiple multifunctional bed frames, allows the bed frame to be disassembled into multiple independent components, facilitating transportation and assembly, while also reducing transportation costs and storage space.

The present invention discloses a VESA board automated production line, including a slide rail tabletop, on which an annular slide rail for conveying workpieces is arranged; a plurality of VESA board positioning jigs are slidably connected to the annular slide rail; the outer side of the slide rail tabletop is provided with a manual VESA board and circular sheet metal placing station, a first VESA board oiling station, etc. in a clockwise direction, and the VESA board positioning jig passes through the above stations in sequence along the conveying direction of the annular slide rail and performs cyclic work. The present invention realizes the integrated integration of multiple equipment and multiple processes, and the annular slide rails connected end to end enable manual labor to be concentratedly distributed in the same section to complete the loading and unloading operations of various parts, and at the same time, with the help of specific equipment for different parts, the assembly of the product can be successfully assembled at one time, reducing the transfer and transportation of workpieces between different equipment and different processes and the storage space, thereby greatly reducing the time, manpower, material and financial costs consumed by the transfer and transportation.

The invention provides a gamma-cyclodextrin crystal form A and a preparation method thereof, an X-ray powder diffraction pattern of the gamma-cyclodextrin crystal form A has characteristic peaks when a 2 theta value is 9.15 + / -0.2 degrees, 12.04 + / -0.2 degrees and 16.35 + / -0.2 degrees, and the preparation method comprises the following steps: adding a certain amount of gamma-cyclodextrin into a certain amount of water, stirring, and carrying out ultrasonic dispersion to obtain a gamma-cyclodextrin crystal form A; the preparation method comprises the following steps: covering a bottle with a cover, heating in a biochemical incubator, dissolving, filtering a dissolved clear solution into a nuclear magnetic tube, adding a certain amount of hydroxypropyl methyl cellulose, vacuumizing, introducing nitrogen, fastening a rubber plug, heating in the biochemical incubator, setting heating and cooling procedures of the biochemical incubator, crystallizing, separating out white cluster-shaped solids at the bottom of the bottle, and drying, thereby obtaining the white cluster-shaped solid. And separating out a solid to obtain the gamma-cyclodextrin crystal form A. The crystal form A is low in energy, more stable, higher in stacking density and convenient to package, transport and store.

This invention discloses a tooling mold for manufacturing brake discs and a method for manufacturing brake discs. The tooling mold includes an insertion section whose shape matches the shape of the heat dissipation holes and achieves an interference fit after insertion. The end of the insertion section is fixedly connected to a sliding groove via a connecting rod. A sliding plate, which cooperates with the sliding groove, enables position adjustment of the insertion section. This invention offers flexible overall adjustment and can be used for combined adjustment of the spacing for brake discs of different sizes, facilitating operator operation. The method of manufacturing brake discs using this invention reduces the difficulty of drilling holes in carbon-carbon composite brake disc preforms, reducing processing volume; it reduces the time and material consumption required for brake disc vapor deposition; it improves the smoothness of the inner surface of the heat dissipation holes, reducing processing volume and avoiding dynamic imbalance and airflow velocity uniformity issues caused by rough holes; it promotes uniform product material, ensures uniform expansion coefficient, and reduces the probability of product cracking.

This invention discloses a hybrid quantization method for image recognition models based on bit-flipping attacks, comprising: Step 1, performing initial quantization on the image recognition model to convert it into an initial quantized model; starting from the first layer of the initial quantized model, performing bit-flipping attacks layer by layer, and recording the sensitivity of each layer under attack; forming a sensitivity set of the initial quantized model from the sensitivities of all layers; constructing an optimization objective function and determining optimization constraints, and determining the optimal hybrid precision quantization strategy for the image recognition model based on the optimization objective function and the optimization constraints; Step 3, setting the bit width for storing the parameters of each layer of the image recognition model according to the optimal hybrid precision quantization strategy, thereby obtaining the optimal hybrid quantization model; Step 4, acquiring the image to be recognized and performing image recognition using the optimal hybrid quantization model.

The embodiment of the present application discloses an image encoding and decoding method, an encoder, a decoder and a storage medium. Before the encoder performs encoding processing according to the MIP mode, the encoder uniformly modifies the initial right shift parameters corresponding to different sizes and different MIP mode numbers according to the offset parameter; wherein the offset parameter is used to indicate the number of right shift bits of the predicted value; and when the encoding processing is performed according to the MIP mode, the encoding processing is performed according to the offset parameter. Before the decoder performs decoding processing according to the MIP mode, the decoder uniformly modifies the initial right shift parameters corresponding to different sizes and different MIP mode numbers according to the offset parameter; wherein the offset parameter is used to indicate the number of right shift bits of the predicted value; and when the decoding processing is performed according to the MIP mode, the decoding processing is performed according to the offset parameter.

The utility model discloses a hard silk screen bamboo fiber double-effect duster cloth which comprises a scrubbing layer which is rough in material and large in friction force, a soft wiping layer which is fine and smooth in material is arranged on the back face of the scrubbing layer, and the scrubbing layer and the soft wiping layer are sewn into a whole in an edge covering and wrapping mode. The scrubbing layer comprises a plurality of flat bamboo hard wires arranged in a grid mode, and the multiple bamboo hard wires are arranged in a crossed mode and connected through first weaving threads in a warp knitting and sewing mode. The soft wiping layer comprises a plurality of U-shaped bamboo fiber lines and second weaving lines which are used for laying and sewing the bamboo fiber lines according to a plane; except the edge covering surface, other materials of the hard silk screen bamboo fiber double-effect cleaning cloth are all made of bamboo fiber materials. Through the combination of the scrubbing layer and the soft wiping layer, one piece of cleaning cloth has double effects, one side can be used for scrubbing stubborn stains, the other side can be used for cleaning and wiping, the cost for purchasing two kitchen and bath sanitary appliances and the placing space are reduced, and the cleaning cloth has the scrubbing function of a steel wire ball.

This utility model relates to a pulp-supported photo frame, including a front support and a rear support, which are detachably connected by a snap-fit ​​structure. Both the front and rear supports have a pulp molding structure. The front support has a protruding window support stamped on the side near the rear support, with a rectangular window formed on the window support. The wall of the window support near the rear support forms a retaining wall, which surrounds the rectangular window. The rear support has a first rectangular ring-shaped reinforcing boss and multiple second rectangular ring-shaped reinforcing bosses stamped on the side away from the front support. The pulp molding structure of both the front and rear supports offers better environmental protection and lower cost compared to traditional plastic products. This application features a protruding window support on the front support and a first rectangular ring-shaped reinforcing boss and multiple second rectangular ring-shaped reinforcing bosses stamped on the rear support. The U-shaped reinforcing grooves form a three-dimensional support network, effectively dispersing stress, increasing structural strength, and ensuring durability, sturdiness, and no deformation during use.

The embodiment of the invention provides an image prediction method, an encoder, a decoder and a storage medium, if a current block uses an MIP mode to determine an intra-frame prediction value of the current block, the encoder sets the value of an MIP mode parameter to indicate to use the MIP mode and writes the value into a code stream; determining an MIP mode of the current block, and determining predicted values of a luminance component and a chrominance component corresponding to the current block according to the MIP mode; and writing the MIP mode of the current block into the code stream. The decoder analyzes the code stream and determines an MIP mode parameter of the current block; if the value of the MIP mode parameter indicates that the current block uses an MIP mode to determine an intra-frame predicted value of the current block, analyzing the code stream, determining the MIP mode of the current block, and determining predicted values of a luminance component and a chrominance component corresponding to the current block according to the MIP mode; and decoding the current block according to the predicted value.

The invention belongs to the technical field of antennas, and relates to a marine rescue inflatable antenna, which comprises a base body part, a radiation part and a feed part, the base body component is of a rotary body structure, and an axial air band gap is formed in the center of the rotary body structure so that the base body component can form a cylindrical structure. A closed cavity is formed in the cylindrical structure and is filled with gas, so that the inflatable antenna for sea rescue is formed; the radiation part is arranged on the side surface of the base body part to radiate an antenna signal; the feed component is connected with the radiation component. According to the invention, the efficiency and success rate of sea rescue can be improved.

The utility model relates to a folding office table which comprises a base frame, a table plate, an installation frame, a hinge rod and a sliding rod. The table plate is connected with the base frame in a turnover mode. The table plate is connected with the base frame through the mounting frame, and a guide groove is formed in the side wall of the mounting frame in the width direction of the mounting frame; one end of the hinge rod is hinged to the base frame, and the other end of the hinge rod is hinged to the mounting frame; one end of the sliding rod is fixedly connected with the base frame, and the other end of the sliding rod extends into the guide groove. According to the foldable office table, the table plate can be folded at the rear end of the base frame when the table plate is not used under the cooperation of the base frame, the table plate, the mounting frame, the hinge rod and the sliding rod, and the storage occupied space of the foldable office table is reduced; and the folding structure is simple, and the phenomenon that the folding office table loses efficacy due to damage of parts in the use process is avoided, so that the service life of the folding office table is effectively prolonged, and the production cost can be effectively reduced.

The embodiment of the utility model discloses a full-storage winder which comprises a circuit board provided with a positioning structure matched with a winding piece of the full-storage winder. The at least two first interface parts are arranged in different areas of the circuit board; the second interface part is arranged on the circuit board; the functional assembly is arranged on the circuit board; and the functional assembly is used for enabling the at least two first interface parts and the second interface part to be electrically connected to work at the same time. According to the circuit board of the full-storage winder, the positioning structure can be arranged on the winding piece and located in the shell, so that the size of the full-storage winder in the storage state can be reduced, the storage size and the occupied storage space of the full-storage winder can be reduced, and the storage effect of the full-storage winder is improved.

The utility model belongs to the technical field of measuring instrument, specifically is a kind of rack corrosion measuring instrument convenient to use, including rack corrosion measuring instrument body, rack corrosion measuring instrument body side wall is equipped with power cord;The end of power cord is equipped with plug;Rack corrosion measuring instrument body surface is fixedly connected with multiple fixed plate;First baffle is rotatably connected in the middle of fixed plate;Second baffle is rotatably connected in the end of first baffle;Third baffle is rotatably connected in the end of second baffle;Screw rod is fixedly connected in the end of rack corrosion measuring instrument body;By increasing first baffle, second baffle and third baffle can be combined to form protection plate when power cord is connected power supply to power supply, can reduce the contact of outside and power cord, increase protection when using, reduce direct contact with power cord, thereby increase the protection of power cord in use, it can reduce storage space after folding, make it more convenient.

The utility model discloses a gas state monitoring sensor calibration device, which relates to the technical field of sensor calibration and comprises a calibrator, a standard micro-water sensor, a standard pressure sensor, a gas circuit adjusting system, a humidity calibration module and a pressure calibration module. The calibrator is provided with a sealing cavity, a first port and a second port, wherein the first port and the second port are communicated with the sealing cavity. According to the technical scheme of the utility model, the humidity calibration module and the pressure calibration module are adopted, so that the dual calibration function of gas humidity and pressure parameters can be realized, the problem of single function of the existing calibration device is solved, multi-parameter integrated calibration is realized, the calibration efficiency is improved, and the application range is widened; by adopting the modular design and switching the four pipelines through the on-off control assembly, a user can select different humidity calibration modules or pressure calibration modules according to needs, the operation is convenient, and the automation level and the operation consistency of the system are improved.

The present application belongs to the technical field of grain fumigation, and particularly relates to a granary medicine feeding machine, which comprises a base plate, a rack fixed on the base plate, a lifting seat in sliding guide cooperation with the rack, a lifting driving mechanism installed on the rack, an outer pipe body vertically downwardly extending through the base plate, an inner pipe body sleeved in the outer pipe body and abutting against the inner bottom end of the outer pipe body, and a connecting mechanism installed on the lifting seat and connected with the top of the inner pipe body. The present application can reduce the resistance of pipe body insertion into the grain pile, reduce the grain occupation of the medicine feeding space and the pipe body blockage, improve the stability of the outer pipe body embedding depth and position, and is beneficial to the release of fumigant in the grain pile according to the predetermined position, so as to realize the fumigation effect.

The utility model belongs to the technical field of concrete formworks, and particularly relates to an M-shaped side wall disassembly-free steel pipe main keel formwork overall moving clamping piece. Comprising an M-shaped clamping piece, two concave included angles are constructed on the M-shaped clamping piece, the two concave included angles are located between two long edges, the M-shaped clamping piece can be clamped on two steel pipe main keels through the two concave included angles, the two concave included angles are separated, a gap for a split screw to penetrate through is reserved between the two steel pipe main keels, the two long edges of the M-shaped clamping piece are connected with a secondary keel, and the two long edges of the M-shaped clamping piece are connected with the secondary keel. The steel pipe main keel and the steel pipe secondary keel are fixedly connected together through a plurality of M-shaped clamping pieces; the steel pipe main keel is fixed on the secondary keel through the M-shaped clamping piece, the main keel and the secondary keel do not need to be dismantled in the formwork moving process and can move along with the formwork, the overall moving area of the formwork is increased, and the installation speed of the formwork is increased.

Embodiments of this application disclose an erasure code-based encoding method and a related device. The method includes: An encoder side obtains a plurality of data blocks in batches, and encodes the plurality of data blocks. In an ith time of encoding, after obtaining an incremental data block, the encoder side performs encoding processing on the incremental data block by using erasure code to generate an incremental parity block, the incremental data block includes at least one data block, and i is an integer greater than 1. The encoder side performs an exclusive OR operation based on the incremental parity block and an original global parity block to generate a first incremental global parity block. The original global parity block is a first incremental global parity block generated by the encoder side in an (i-1)th time of encoding, and an original data block is a data block obtained by the encoder side before the (i-1)th time of encoding. The first incremental global parity block is used to reconstruct a damaged data block in the original data block and the incremental data block. Encoding flexibility is significantly increased, the encoding processing at the encoder side is averaged, and a computing hotspot problem caused by full encoding is avoided.

Embodiments of the present specification disclose a data tracing method and device, a storage medium and equipment. The method comprises: parsing a to-be-traced structured query language (SQL), to obtain an abstract syntax tree corresponding to the to-be-traced SQL; based on the abstract syntax tree, obtaining predetermined element information corresponding to the to-be-traced SQL; based on the predetermined element information and the syntax structure of the to-be-traced SQL, constructing a reverse SQL corresponding to the to-be-traced SQL; and performing data tracing processing based on the reverse SQL.