42results about How to "Efficient output" patented technology

The embodiment of the application discloses a kind of based on wavefront parallel processing to the method, device, storage medium and computer equipment of image H.264 encoding, it is related to image processing field.The application is processed adjacent macroblock line by double encoder parallelly, and respectively first buffer and second buffer are written to encoding bit stream.In macroblock line encoding completion, first buffer remaining bit stream and entropy encoding remaining bit stream are merged in real time, if the length after merging is less than 8 integer multiples, then appropriate bit stream is dynamically taken from the head of second buffer and spliced, complete code stream in line with H.264 standard is generated and outputed.After encoding is completed, the minimum splicing length is used as the offset of second buffer, continue to output byte stream, and the encoding of next macroblock line is started.The application only needs double macroblock line buffer, significantly improves the encoding speed, reduces memory occupation, and does not need to modify decoder, and compatibility is strong.

The application discloses a kind of man-machine collaborative multi-modal data intelligent labeling and quality checking method, it is related to man-machine collaborative technical field.The method first obtains at least containing two modalities of original data of image, text, audio, generates multi-modal joint data sample after pre-processing and modal alignment;Input pre-trained multi-modal AI pre-labeling model generates pre-labeling result with confidence;Through cross-modal consistency check calculation consistency score, sample is divided into high, medium, low confidence three categories;High confidence sample is directly adopted and enters the sampling pool, and medium and low confidence sample is based on labeling personnel ability portrait and AI model dynamic capability evaluation and carries out intelligent task distribution.The application effectively solves the problem that efficiency and quality are difficult to balance in multi-modal data labeling, significantly improves labeling accuracy and automation rate, reduces labeling cost, and can meet the construction demand of large-scale high-quality multi-modal data set.

An electrochemical device and an electronic device comprising the same. The electrochemical device comprises at least one bipolar current collector (10) sealedly connected with an outer package (20), forming two independent sealed cavities on both sides of the bipolar current collector (10), each of which comprises an electrode assembly (30) and an electrolyte, one side of the bipolar current collector (10) is electrically connected with the outermost positive electrode tab of the adjacent electrode assembly (30), and the other side of the bipolar current collector (10) is electrically connected with the outermost negative electrode tab of the adjacent electrode assembly (30). Through the introduction of the bipolar current collector (10) and the sealed design of the bipolar current collector (10) and the inner layer of the outer package (20), the ion insulation between the multiple electrode assemblies (30) is realized, and the reliability of the high-output voltage battery and the effective output of the electric energy are realized.

The invention discloses a lightweight image segmentation method and system based on NeuralODE neural network architecture, and the method comprises the steps: obtaining to-be-segmented image data, extracting an initial feature map through an initial convolution layer, and generating an intermediate feature map through a separable convolution layer; inputting the intermediate feature map into an enhanced Shenchang differential equation block, executing multi-step continuous feature evolution and introducing preorder output feature weighted fusion in each step to obtain an evolved feature map; generating a main feature map and a derived feature map, splicing the main feature map and the derived feature map in a channel dimension, and outputting a compressed feature map; a decoupling knowledge distillation strategy is adopted, teacher model probability distribution is decoupled into two types of probability distribution, corresponding loss is calculated, and total distillation loss is constructed to adjust network parameters; and inputting a feature map output by the trained network model into a segmentation prediction head, and executing pixel-level prediction operation on the feature map through the segmentation prediction head to generate a segmentation result of the to-be-segmented image. The method improves the segmentation precision, and gives consideration to the deployment efficiency and the segmentation performance.

The invention relates to a photoelectric sensing element, a photoelectric sensing array and a preparation method, the photoelectric sensing element comprises a substrate, a transistor module and a photoelectric detection module, the transistor module comprises a grid electrode, a first dielectric layer, an active layer, a first source electrode, a first drain electrode and a first passivation layer; the photoelectric detection module is sequentially provided with a second dielectric layer, a second source electrode, a second drain electrode and a second passivation layer from bottom to top; the first source electrode, the first drain electrode, the second drain electrode, the second source electrode and an external reading circuit are connected in series to form a loop; and a photosensitive induction layer of the photoelectric detection module is filled into a channel between the second source electrode and the second drain electrode. The transistor module and the photoelectric detection module are integrated on the substrate and are connected in series to form a complete loop, and the photosensitive sensing layer made of the two-dimensional transition metal chalcogenide is adopted, so that efficient capture, conversion and output of optical signals are realized, the sensitivity and response speed of optical signal detection are improved, and the structural stability of the device and the accuracy of signal transmission are guaranteed.

The invention discloses a starting method and device of an image processing system, electronic equipment and a computer readable storage medium, and the image processing system comprises a system processor, an image sensor, an image processor, an image signal expander, an image memory and a display. The system processor is respectively connected with the image sensor, the image processor, the image memory and the image signal expander, the image sensor is connected with the image processor, the image processor is connected with the image signal expander, and the image signal expander is respectively connected with the image memory and the display. The starting method of the image processing system comprises the following steps: judging whether signal interaction of the image processing system in a starting process is abnormal or not; and when it is determined that the signal interaction is abnormal, the image signal expander is restarted through the system processor. By applying the technical scheme provided by the invention, the problem of starting failure of the image processing system can be quickly solved, and the system starting time is shortened.

The present application relates to the technical field of underwear gluing image recognition, and particularly relates to a kind of underwear gluing image recognition method and system based on high-precision positioning, the method comprises the following steps: obtaining no trace underwear image data;The no trace underwear image data is divided into multiple local regions, and local brightness feature information is obtained;According to the local brightness feature information of each local region, the edge detection threshold of each local region is determined;Edge detection operator is applied to each multi-resolution image, and the edge graph of each multi-resolution image is extracted in combination with the edge detection threshold of each local region;Fusion each edge graph, and obtain preliminary edge contour;The preliminary edge contour is subjected to morphological processing, smoothing processing and connection processing, and refined gluing contour is obtained;The refined gluing contour is converted into coordinate point sequence, and the coordinate point sequence is output as gluing path instruction.The above is helpful to improve product quality and production efficiency.

This application relates to the field of medical data processing technology and discloses a method and system for screening infertile patients with polycystic ovary syndrome (PCOS). The method includes: acquiring data on the number of antral follicles, anti-Müllerian hormone (AMH) concentration, and ovarian volume from multiple patient monitoring cycles; performing feature extraction and normalization to obtain a set of normalized feature vectors; determining the synchronous relationship between the number of antral follicles and ovarian volume based on the vector set, and identifying signs of asynchronous changes; combining hormone concentrations for joint grouping to obtain the distribution of abnormal indicator clusters and identifying ovarian function stage markers; performing sequence matching based on the stage markers of continuous cycles and historical medical records to generate an ovarian function progression trajectory curve; extracting inflection point features from the curve and performing causal association rule mining to construct a logical chain between abnormal indicators and clinical outcomes; and determining conditions layer by layer according to the logical chain to output the final screening results. This application improves the accuracy of screening infertile patients with PCOS.

The invention relates to the technical field of semiconductors, and provides an artificial intelligence assisted czochralski method crystal seeding process method. According to the method, seeding process parameters directly related to crystal growth are selected from seeding historical data, and part of seeding historical data are selected for calibration to obtain the deviation between the actual seeding temperature and the ideal seeding temperature so as to obtain the ideal seeding temperature under the corresponding seeding process parameters, so that an artificial intelligence supervised learning model is trained; the trained artificial intelligence supervised learning model is adopted, the ideal seeding temperature of the furnace is obtained through prediction of the artificial intelligence supervised learning model according to the current seeding process parameters, a process engineer is assisted to set the proper seeding temperature, and convenience is provided for production; deviation is automatically corrected along with continuous production, and monocrystalline silicon crystals are stably, efficiently and continuously produced; according to the method, the defects of dependence on artificial experience, low efficiency, non-standardized operation and the like in manual seeding temperature determination can be overcome, and the method can be suitable for modernized large-scale production.

The utility model relates to the technical field of fan motors, and provides an exhaust fan motor for an oven, which comprises a cross-flow fan, one side of the cross-flow fan is provided with a motor, the motor is used for providing power for the motor, the other side of the cross-flow fan is provided with a centrifugal fan, the cross-flow fan comprises a centrifugal fan impeller, and the centrifugal fan impeller is used for providing power for the centrifugal fan. A connecting shaft is installed on one side of an impeller of the centrifugal fan, the connecting shaft is in transmission connection with the centrifugal fan, the cross-flow fan comprises a first machine shell, the first machine shell is provided with a first air outlet, the centrifugal fan comprises a second machine shell, and the second machine shell is provided with a second air outlet. In the utility model, the cross-flow fan and the centrifugal fan are simultaneously driven by the single motor, so that the equipment can realize high-efficiency airflow output in different directions, and the requirement of multi-direction ventilation is met. The integrated design not only improves the flexibility of equipment use, but also simplifies the installation and configuration processes, and reduces the demand for additional equipment.

The application discloses a kind of low-carbon scheduling method, system, equipment and medium of power system, belong to power grid low-carbon scheduling technical field, method includes initialization parameter and constructs collaborative space;By particle swarm module optimization reinforcement learning hyperparameter;Reinforcement learning module interactive experience data generation;Extract priority sample and elite particle realize two-way experience sharing;Converge after output scheduling strategy.System includes initialization module, collaborative space construction module, hyperparameter optimization module, interactive storage module, two-way experience sharing module and strategy output module.The application is coupled by constructing dynamic collaborative optimization space and two-way experience sharing mechanism, particle swarm and deep reinforcement learning.Feedback guides particle swarm to realize adaptive disturbance, avoid local optimum trap.At the same time, spontaneous optimization network parameter, significantly enhance the adaptability and generalization ability of algorithm.The scheme breaks through single algorithm decision bottleneck, improves low-carbon scheduling efficiency and power system reliability.

The application relates to a physical constraint electron microscope simulation generation and segmentation method and system, and relates to the technical field of computer vision and electron microscopic analysis. The method first adopts a simulation training data generation engine to construct a composite atomic structure model subject to physical constraints, synchronously generates a strictly aligned pixel-level segmentation mask, performs HRTEM imaging simulation and image enhancement based on the composite atomic structure model, establishes a simulation dataset in combination with the pixel-level segmentation mask, and establishes a real dataset based on real HRTEM imaging. Then, a simulation-to-real training process is performed, the segmentation model is pre-trained using the simulation dataset, and fine tuning is performed using the real dataset until the model training is completed. The model is applied to segmentation of input data. Without relying on a large number of high-quality pixel-level labels, the segmentation accuracy and training robustness of the segmentation model are effectively improved, the analysis efficiency of the segmentation model is significantly improved, and high-throughput and quasi-real-time applications are supported.

The utility model relates to a kind of planet carrier output shaft in planetary reducer technical field, including shaft body and several gears, the one end of shaft body is provided with several through grooves along the axis, several gears are detachably mounted in several through grooves by quick-release assembly, the other end of shaft body is equipped with connecting rod.The utility model is through quick-release assembly, gear is installed in the through groove on shaft body, to realize the quick assembly and disassembly of gear, improve assembly efficiency and maintenance convenience, to reduce maintenance time, improve the operating efficiency of equipment.And by connecting rod and external driving element cooperation, to realize the efficient transmission and stable output of power.

This application relates to the field of hyperspectral remote sensing application technology, and provides a hyperspectral methane detection method based on spatial-spectral joint analysis. First, a sample dataset is constructed, and simultaneously, a methane detection neural network is built to extract the spatial-spectral features of the methane plume from the sample dataset. Then, a preprocessing module based on fundamental laws is established to form a hard constraint mechanism, outputting standardized spatial-spectral features of the methane plume to obtain a predicted methane concentration distribution map. A physical information penalty term is constructed to form a soft physical constraint, minimizing the error between the predicted methane concentration distribution map and the pseudo-true label, thus incentivizing the network to generate a full-resolution methane concentration distribution map. Through a deeply fused spatial-spectral joint strategy, relying on the synergistic effect of the hard constraint mechanism and the soft physical constraint, effective suppression of background noise can be achieved in complex atmospheric environments and various background interference scenarios, significantly improving the accuracy and reliability of signal detection. The detection efficiency is high, making it suitable for large-scale and efficient processing of massive hyperspectral data.

The invention relates to the technical field of vehicle intelligent control, and discloses a vehicle braking control method and system, computer equipment and a storage medium. The method comprises the steps that in response to a braking request signal, the target braking pressure of a target vehicle is determined; determining a target control strategy of the brake motor based on the target brake pressure and the current state data of the brake motor; and based on the target control strategy, a brake motor is controlled to drive a brake piston to move, so that brake control over the target vehicle is achieved. And in response to the braking request signal, the target braking pressure is determined, a control strategy is formulated in combination with the motor state data, and an accurate plan of braking action is achieved. And finally, the brake motor is controlled to drive the brake piston to move, the limitation of traditional one-way pressure building is broken through, two-way pressure building and synchronous liquid supplementing are achieved, and the problem of insufficient brake liquid during continuous braking is solved. And the braking stability and comfort are enhanced while the cost advantage is guaranteed.

The invention relates to the technical field of pipeline design, and discloses a large-scale water delivery pipeline deepening design method based on a visual programming language, and the method comprises the steps: extracting key data from a design drawing of a water delivery pipeline, and carrying out the preprocessing to obtain standard key data; determining spatial information of the plurality of control points based on the standard key data, and constructing a three-dimensional pipe section model according to the spatial information; preset section parameters are obtained, a three-dimensional entity pipeline model is generated in combination with the three-dimensional pipe section model, and key nodes are added for the three-dimensional entity pipeline model in combination with preset key node parameters; and engineering drawings and part expansion drawings of the components and the parts are generated based on the numbers. Through the cross-platform data engine, full-link automatic processing of design drawing-three-dimensional model-construction drawing-manufacturing drawing data is achieved, the parameterization degree of pipeline design is improved, the deepening design efficiency and precision are improved, and quick response can be achieved by means of the architecture of standard key data and a parameterization model in the face of design change.

The invention discloses a method for extracting prenuclide and preparing medical alpha isotope on a large scale based on monazite discharged waste slag. According to the method, high-value conversion of radioactive waste is realized by comprehensively utilizing the monazite waste slag. The alpha isotope is used as a core raw material for extracting preposed nuclide, so that the environmental harm of radioactive waste residues is reduced, waste is turned into wealth, and a stable raw material source is provided for preparing the medical alpha isotope. A complete technical chain for extracting the medical alpha isotope 225Ac / 212Pb from the monazite waste slag containing 226Ra and 228Ra is constructed, and efficient output of high-value nuclides is achieved. 226Ra and 228Ra in the discharged waste slag are obtained through a step-by-step extraction process flow of a coarse radium product and a fine radium product, a radium target of an ECANS device is manufactured, and hot medical alpha isotopes such as 225Ac, 212Pb and the like can be prepared on a large scale by utilizing a photo-nuclear reaction. According to the preparation mode, the utilization rate of the discharged waste slag is increased, and the requirement of radioactive diagnosis and treatment on the medical alpha isotope can be met.

The utility model is suitable for solar technology field provides a kind of composite photovoltaic module, it includes photovoltaic module, light and heat module and frame, photovoltaic module includes perovskite battery group and the crystalline silicon battery group being arranged below perovskite battery group.Light and heat module includes upper air flow channel, lower air flow channel, heat collecting plate, foam metal layer and heat preservation layer.Frame is arranged around the outer periphery of photovoltaic module and light and heat module, first aperture and second aperture are set on frame, first aperture is communicated with cold air inlet, and second aperture is communicated with hot air outlet.Photovoltaic module is composed of perovskite battery group and crystalline silicon battery group, perovskite battery group and crystalline silicon battery group can absorb different wavelength light intensity, with the advantage of improving solar full spectrum utilization rate, because the double air flow channel and heat preservation measure setting of light and heat module setting, so solve the problem of low photovoltaic module heat energy utilization rate and water pipe freezing in severe cold area.

This invention belongs to the interdisciplinary field of computer vision and medical health monitoring, and discloses a non-contact heart rate measurement method and system based on spatiotemporal augmentation networks, solving the problems of inconvenience of traditional contact measurement and weak anti-interference of existing rPPG technology. The method includes: performing illumination normalization on the original facial video frames and constructing an input tensor; inputting an end-to-end spatiotemporal augmentation neural network (including 3D central difference convolution and multi-dimensional attention module) to reconstruct the rPPG time-series signal; and converting the heart rate through bandpass filtering and fast Fourier transform. This scheme achieves mean absolute errors of 0.24 and 0.40 on the UBFC-rPPG and PURE datasets, respectively, is resistant to illumination / motion interference, is highly efficient end-to-end, has only 0.99M network parameters, is compatible with servers (inference ≥35FPS) and embedded devices (inference ≥28FPS), and has excellent generalization performance.

The invention belongs to the technical field of unmanned aerial vehicles, and particularly relates to an omnidirectional flight unmanned aerial vehicle based on a passive hinge type vector power system. Comprising a machine body, vector power systems and machine arms, the three machine arms are arranged on the machine body in the circumferential direction at equal intervals, and the tail ends of the three machine arms are connected with the three vector power systems correspondingly; the vector power system comprises a power system fixing frame, a two-shaft connecting piece and a plurality of power rotors, one end of the two-shaft connecting piece is rotationally connected with the vehicle arm, the other end of the two-shaft connecting piece is rotationally connected with the power system fixing frame, and the power rotors are arranged on the power system fixing frame. The power system fixing frame is of an H-shaped structure, and four power rotors are arranged at the four ends of the power system fixing frame correspondingly. The whole power vector is changed under the condition that no extra actuator is added, hovering in the air can be achieved in any posture, and omni-directional movement is achieved under the condition that any posture is kept.

The application relates to the technical field of high-efficiency hot forging, and discloses a large-tonnage high-efficiency hot forging equipment which comprises an oil pressure device for providing power output and an oil pipe loop system for realizing oil circuit control and energy management. The large-tonnage high-efficiency hot forging equipment realizes energy storage and recycling through cooperation of an accumulator group and a variable oil pump group in the oil pipe loop system. In an initial stage, the variable oil pump group charges the accumulator to reserve energy, and in a return stroke stage, the accumulator group recycles energy through the oil pump. Compared with a traditional single main oil cylinder driving equipment, the energy consumption is greatly reduced, the problem that power demand increases rapidly with the increase of tonnage and leads to a sharp increase of power grid load is solved, the quick cylinder in the oil pipe loop system is controlled by an independent cartridge valve, the oil circuit is integrated with a two-way cartridge valve to realize millisecond-level response, the quick cylinder can complete the air travel drop in a short time, the accumulator group is connected to the main oil cylinder inlet through high-pressure hoses in parallel, and the volume of oil liquid is instantaneously released to push the main oil cylinder, so that the system reaches large-tonnage output.

This invention relates to a modular electric drive axle system and a vehicle. The system includes a main housing, a first housing, a second housing, a differential, a first drive module, and a second drive module. The first housing and the second housing are detachably connected to the left and right sides of the main housing, respectively. The differential is disposed within the main housing. The first drive module is disposed within the first housing and includes a first motor and a first transmission system connected to the first motor. The second drive module is disposed within the second housing and includes a second motor and a second transmission system connected to the second motor. The first transmission system and the second transmission system are respectively connected to the differential. The first drive module and the second drive module are structurally independent and can independently cooperate with the main housing to form a single-motor electric drive axle system, or cooperate together with the main housing to form a dual-motor electric drive axle system.

The utility model relates to the technical field of aerated brick production, and discloses a raw material stirring device for aerated brick production, which comprises a double-layer frame, the top of the double-layer frame is fixedly connected with a stirring barrel, a stirring mechanism is arranged in the stirring barrel, the stirring mechanism is used for stirring and mixing raw materials, and a discharging mechanism is arranged at the bottom of the inner side of the double-layer frame. The discharging mechanism is used for improving discharging convenience, the bottom end of the stirring barrel communicates with an electromagnetic connecting valve, and the stirring barrel communicates with the discharging mechanism through the electromagnetic connecting valve. The cover plate is in threaded connection with the stirring barrel to form sealing, the stirring motor drives the stirring shaft to drive the cross-shaped stirring impeller to rotate in the impeller groove, the upper impeller and the lower impeller stir in a layered mode to generate reverse pressure to promote three-dimensional convection of raw materials, and in the aerated brick production process, uniform mixing is achieved through layered stirring and three-dimensional convection, and discharging is accurately controlled. The raw material stirring efficiency and the mixing quality are improved.

This invention relates to the field of lithium battery production control technology, and discloses a production control method for achieving efficient production of lithium battery cells. The method acquires batch characteristic data of lithium battery cells, including electrode material ratio characteristics and electrolyte composition characteristics; extracts a corresponding set of process parameter configurations based on the batch characteristic data; identifies key control nodes in the lithium battery production process based on the set of process parameter configurations, including material ratio adjustment nodes and process parameter switching nodes; calculates parameter deviation data in the production process based on the key control nodes, including cumulative deviations within the same batch and parameter jump values ​​across batches; and extracts a set of reusable process control parameters based on the parameter deviation data, including general process parameters and specific process parameters.

This invention discloses an elevator counterweight frame, comprising a frame, a sheave, and a counterweight block. The frame has a space for accommodating the counterweight block. At least one sheave is provided, mounted on the counterweight upper beam at the top of the frame and rotatably connected to it. A power generation module is located on one side of the sheave body. The power generation module includes a stator frame, which is fitted onto and rotatably connected to the sheave's axle. Multiple winding posts are evenly distributed around the outer side wall of the stator frame, with coils wound around them. Multiple permanent magnets are evenly distributed around the edge of the sheave body, corresponding to the winding posts. Strain gauges are mounted on the counterweight upper beam. In this invention, the strain gauges and the power generation module work together to monitor the weight of the counterweight block, while the power generation module generates its own power to supply the strain gauges. Throughout the process, the mechanical energy of the sheave is converted into electrical energy, achieving self-sufficiency in electrical energy without the need for an external power source.

This invention discloses an infrared acupoint moxibustion device, comprising: a moxibustion head, one end of which is proximal to the skin and the other end of which is distal to the skin; the proximal end of the moxibustion head has a lens and an external air inlet, and the distal end of the moxibustion head, or the side wall near the distal end, has an air outlet and a wiring terminal; the moxibustion head internally comprises a support base, a reflector, a cooling fan, and an infrared heating element; the support base internally has a cavity, and the support base has a front support surface and a rear support surface, the front support surface having a front opening and an internal air inlet, and the rear support surface having a rear opening; the reflector is mounted on the front support surface of the support base; the infrared heating element is disposed inside the front opening and electrically connected to the wiring terminal; the cooling fan is mounted at the rear port. This invention ensures efficient heat dissipation, avoids burns, improves equipment safety and service life, and optimizes the focusing and output of infrared energy, making moxibustion more precise, safe, and stable.

The embodiment of the application provides a kind of acquisition integrated assembly and battery pack, acquisition integrated assembly includes insulating plate, short-circuit row, output row and acquisition piece, insulating plate is integrally provided with first mounting portion, second mounting portion and third mounting portion;Short-circuit row is used for bridging the positive pole and the negative pole of two batteries, short-circuit row is fixed to the first mounting portion, short-circuit row has guide hole, and the guide hole is clamped with the positioning element of the first mounting portion;Output row is used for outputting the electric energy of battery module, and output row is fixed to the second mounting portion;Acquisition piece is used for collecting the signal of battery module, and acquisition piece is fixed to the third mounting portion.The application is positioned and fixed to the insulating row by short-circuit row, output row and acquisition piece quickly, improves the assembly processing efficiency and working yield of acquisition integrated assembly, and finally guarantees the normal stable work of battery pack.

The invention discloses an improved C-dump converter energy and signal simultaneous transmission strategy. In order to solve the problem that energy in a follow current stage of a traditional C-dump converter can only be fed back to an excitation power supply and power generation energy output cannot be realized, a follow current loop suitable for a power generation state is constructed by adjusting the connection position of an energy storage capacitor in a circuit, so that the control independence of each phase is maintained and the use of devices is not increased; and effective output of power generation energy is realized. The flyback converter is adopted to widen the voltage output range and reduce the output voltage fluctuation at the same time. In order to realize synchronous transmission of energy and signals, a frequency shift keying modulation method is used, and ripple signals containing communication data are embedded in output voltage; and a receiving end performs information demodulation by using fast Fourier transform, so that an information transmission system is simplified. According to the scheme, the system complexity is remarkably reduced, and the output electric energy quality of the switched reluctance generator is effectively improved.

The application discloses a multifunctional stirring conveyor, and belongs to the technical field of multifunctional stirring conveyors.The multifunctional stirring conveyor comprises a stirring barrel, a stirring device for stirring raw materials (in slurry form) is arranged in the stirring barrel, the stirring device comprises a supporting plate, the supporting plate is fixedly connected to the top of the stirring barrel, a first rotating rod is rotatably connected to the bottom of the supporting plate through a bearing, a limiting block is fixedly connected to the outer wall of the first rotating rod, and a stirring rod is slidably connected to the outer wall of the limiting block.The stirring rod is moved up and down to push the raw materials (in slurry form), so that the raw materials (in slurry form) are prevented from depositing at the bottom of the stirring device, the stirring is ensured to be more uniform, the first circular gear and the second circular gear interact to transmit power, the stirring blades are rotated, and the raw materials (in slurry form) are more comprehensively stirred.The problem of dead angles that is prone to occurring during stirring is solved, and the raw materials (in slurry form) are more uniformly mixed.

The invention relates to the technical field of 3D Gaussian splashing, in particular to a three-dimensional Gaussian compression method and system based on just perceptible distortion, and the method comprises the steps: collecting images, integrating and building a data set, and inputting the data set into a JND model to obtain a corresponding JND image; inputting the data set and the corresponding JND image into an anchor point-based 3DGS compression method model for training, sequentially obtaining the opacity and the JND score accumulation value of the anchor point, designing a loss function, and optimizing the attribute of the anchor point; judging whether the anchor point is pruned or not according to the opacity and the JND score accumulated value, and completing training of a 3DGS compression method model based on the anchor point; and the trained anchor point-based 3DGS compression method model outputs a rendered image. Extreme compression is achieved on the premise that human visual perception is not affected based on just perceptible differences, and an effective technical scheme is provided for storage and transmission of three-dimensional visual data.