210results about How to "Reduce the difference" patented technology

The invention relates to the technical field of battery SOC equalization, and particularly discloses an energy storage system battery SOC equalization method based on an EMS, and the method comprises the steps: determining SOC equalization influence factors of each battery based on the EMS, analyzing the SOC real-time use state of each battery, obtaining the SOC equalization state characteristic value of each battery in combination with the SOC equalization influence factors of each battery, and determining the SOC equalization state judgment value of each battery. According to the method, whether the battery needs to be subjected to SOC equalization or not is judged, an SOC equalization mode is determined, the temperature state of the battery SOC equalization process is monitored, whether the battery SOC equalization process is overheated or not is judged, and the problems that a traditional equalization method is low in response speed when coping with a scene with frequent load change and high-power requirements, a fixed SOC equalization threshold value is set, and the battery SOC equalization process is overheated or not are solved. The method solves the problem that the prior art is poor in adaptability to dynamic working conditions such as load fluctuation, unnecessary energy consumption is avoided, the system safety can be improved, and the probability of sudden failures is reduced.

The invention discloses a deep tunnel surrounding rock mechanical parameter inversion method based on a three-dimensional brittle failure zone contour and sparrow optimization algorithm-information vector machine (SSA-IVM) combined optimization algorithm. The engineering technical problem that due to the fact that excavation instantaneous displacement is difficult to monitor, deep tunnel surrounding rock mechanical parameters are difficult to reasonably obtain through displacement back analysis is solved. The method comprises the following steps: firstly, constructing a tunnel FLAC3D numerical simulation model with the same ground stress condition at the occurrence position of a three-dimensional brittle failure zone; secondly, taking an absolute error between the total number of computational grid units in the actually measured brittle failure zone and the total number of computational grid units entering a plastic state in the range of the actually measured brittle failure zone after calculation of the FLAC3D numerical model as an optimization objective function; and then, by taking the tunnel surrounding rock mechanical parameters as optimization variables and taking a target function reaching a global minimum value as a target, performing global optimization by combining a tunnel FLAC3D numerical model and adopting an SSA-IVM joint optimization algorithm, thereby obtaining reasonable tunnel surrounding rock mechanical parameters.

The invention discloses a graphic design method for balancing the copper thickness of an impedance line. The graphic design method comprises the following steps: S1, determining a first residual copper rate of a graphic area of an impedance strip layer on a circuit board; s2, acquiring a second residual copper rate of an impedance strip area of an impedance strip layer on the circuit board; s3, the absolute value m of the difference between the second residual copper rate and the first residual copper rate is calculated; s4, m is compared with 5%, and if m is larger than 5%, the step S41 is executed; if m is less than or equal to 5%, executing the step S42; s41, adjusting the width of the auxiliary copper sheet on the two sides of the impedance line in the impedance strip area, and returning to execute the step S2, the step S3 and the step S4 after the adjustment is completed until m is less than or equal to 5%; and S42, simulating a circuit of the pattern area on the auxiliary copper sheet to carry out dummy impedance line distribution in the impedance strip. According to the method, the difference between the copper plating thickness of the impedance strip coupon impedance line of an FCBGA product and the copper plating thickness of an actual graphic circuit can be reduced, and it is guaranteed that the impedance value simulated by the coupon impedance line of the impedance strip has reference significance.

The invention discloses a V2X cooperative sensing method and system based on a lightweight domain adapter. According to the invention, a lightweight domain adapter network is placed in a feature map input branch of each cooperative vehicle of a first V2X cooperative sensing network carried by an own vehicle to form a second V2X cooperative sensing network, and cooperative sensing of the own vehicle and the cooperative vehicles is realized by using the trained second V2X cooperative sensing network. According to the invention, by introducing the domain adapter network, collaborative 3D target detection under the condition of heterogeneous feature maps of the own vehicle and the collaborative vehicle is successfully realized. According to the method, under the condition that the whole basic cooperative detection network does not need to be retrained, the domain adapters corresponding to different feature maps are inserted, so that the domain difference between heterogeneous feature maps is effectively reduced, the detection performance and scene adaptability of cooperative 3D target detection based on V2X are greatly improved, and the detection efficiency is improved. The method has a high practical value for improving the collaborative perception performance of automatic driving.

The invention provides a model training method, an industrial internet anomaly detection method and a related device, and relates to the field of industrial internet. The model training method comprises the following steps: acquiring a training sample comprising a plurality of time sequence data samples; inputting the training sample into a domain projection module to perform feature extraction of a unified dimension to obtain a plurality of target features of a target dimension; inputting the plurality of target features into a feature mixing enhancement module for feature mixing enhancement processing to obtain positive features and negative features of a plurality of target dimensions; inputting the target features, the positive features and the negative features into an anomaly detection module for anomaly detection processing to obtain inter-block characterization and intra-block characterization corresponding to each feature, and characterization loss between the inter-block characterization and the intra-block characterization; and adjusting parameters of the anomaly detection model based on a first alignment loss among the plurality of target features, a second alignment loss among the plurality of in-block representations and a target loss determined by the representation loss, and obtaining the anomaly detection model with relatively good generalization ability.

The invention relates to a method for controlling the axial resistivity of a crystal bar and the crystal bar, and the method comprises the steps: carrying out the equal-diameter growth of a first crystal bar according to a preset equal-diameter parameter, measuring the axial resistivity of the first crystal bar, and drawing a change curve of the axial resistivity of the first crystal bar; calculating the unevenness of the axial resistivity of the first crystal bar; when the unevenness of the axial resistivity of the first crystal bar is greater than or equal to 17%, the evaporation of the antimony element is promoted or inhibited by adjusting the pressure parameter aiming at different change stages of the axial resistivity, so that the concentration of the antimony element in the silicon solution is adjusted; the risk that the reduction trend of the axial resistivity of the crystal bar is large due to the fact that the concentration of the antimony element in the silicon solution is high is reduced, and the risk that the increase trend of the axial resistivity of the crystal bar is large due to the fact that the concentration of the antimony element in the silicon solution is low is also reduced, namely, the difference value between the maximum resistivity and the minimum resistivity is reduced; and thus, the nonuniformity of the axial resistivity is reduced.

The invention relates to the technical field of product glue dispensing, in particular to a glue amount dynamic balancing method for a multi-valve glue dispenser and the multi-valve glue dispenser. The method comprises the steps that S1, glue cylinders of all the dispensing valves are filled with glue; s2, the object to be subjected to dispensing operation is moved to the operation platform deck, dispensing area division is conducted on the object to be subjected to dispensing operation according to the arrangement positions of the dispensing valves and the number of the dispensing valves in combination with the remaining amount of glue in the glue barrel, and the number of the dispensing areas corresponds to the number of the dispensing valves one to one; s3, each dispensing valve is controlled to carry out dispensing operation according to the corresponding dispensing area; s4, obtaining the residual amount of glue in the glue cylinder of each glue dispensing valve after the current glue dispensing operation is completed; and S5, repeating the steps S2 to S4, and returning to the step S1 when the glue allowance in the glue cylinder of each dispensing valve is in a relatively balanced state and any one dispensing valve cannot meet the dispensing requirement. The glue in the glue cylinder of each dispensing valve can be dynamically balanced, the frequency of frequent shutdown is reduced, and the dispensing production efficiency is improved.

The embodiment of the invention discloses a three-dimensional reconstruction method and device based on sparse view inline prior. A specific embodiment of the method comprises the following steps: determining a sparse point cloud of a three-dimensional scene from a sparse view angle image; according to the sparse point cloud, initializing the three-dimensional Gaussian point cloud to obtain an initialized three-dimensional Gaussian point cloud; performing random sampling in a preset range of each sparse view angle in each sparse view angle of the sparse view angle image so as to generate a pseudo view angle and obtain a pseudo view angle set; determining a plurality of optimization regular terms according to each sparse view angle and pseudo view angle set; determining a loss function according to the plurality of optimization regular terms; and according to the loss function, optimizing parameters of the initialized three-dimensional Gaussian point cloud so as to carry out sparse visual angle three-dimensional reconstruction. According to the embodiment, mode offset is inhibited, and the three-dimensional reconstruction quality under the sparse view angle is improved.

The invention relates to the technical field of natural language processing, in particular to a cross-domain sentiment analysis method fusing comparative learning and a domain alignment strategy, which comprises the following steps: inputting source domain and target domain texts into a model encoder to obtain source domain and target domain text features, and outputting each text feature through a model decoder to correspondingly generate a tag; constructing contrast loss, domain adversarial network loss and regularization item loss based on contrast learning, the domain adversarial network and a KL divergence adjustment mechanism; a pseudo label prediction model is obtained by minimizing the first loss function training model; inputting the source domain label and the target domain label into a pseudo-label prediction model encoder to obtain source domain label characteristics and target domain label characteristics, and outputting enhanced texts corresponding to the label characteristics through a pseudo-label prediction model decoder; constructing node matching and structure matching loss according to a graph matching principle; and training the pseudo label prediction model by minimizing the second loss function to obtain an enhanced text generation model. According to the invention, the cross-domain sentiment analysis effect is improved.

The invention belongs to the technical field of highlight injection mold temperature controller data regulation and control, and particularly discloses and provides a highlight injection mold temperature controller intelligent control regulation and control system which comprises the steps that whether an injection mold insert is replaced or not is recognized, flaw detection is conducted on a pre-material, and operation data of a highlight injection mold temperature controller are preset; the defect of the pre-material can be prevented from being amplified in the injection molding process; by identifying the defect parameters corresponding to the operation data, the specific operation parameters causing the defects can be accurately identified, so that the preset operation data of the next operation of the highlight injection mold temperature controller can be corrected in a targeted manner, the generation of the defects can be reduced, the consistency of the quality of products produced by injection molding each time is ensured, and the production efficiency is improved. The difference between products is reduced; the regulation and control trend of the preset operation data is determined by identifying the operation parameter deviation trend of the highlight injection molding target, and a scientific basis can be provided for a decision maker, so that the decision maker can make a more accurate and reasonable decision.

Provided is an optical control element capable of arranging an input portion and an output portion of an optical waveguide on the same side of a substrate in which the optical waveguide is formed, and capable of reducing differences in the configuration of each branch waveguide while minimizing optical path differences between branch waveguides. The optical control element has a substrate (1) having an electro-optic effect, an optical waveguide (2) formed in the substrate, and a control electrode that controls an optical wave propagating in the optical waveguide, characterized in that an input portion (incident light (L1)) and an output portion (emitted light (L2)) of the optical waveguide are formed on the same side of the substrate, the optical waveguide has at least one Mach-Zehnder type optical waveguide portion (MZ) having two branch waveguides (21, 22) branched from one optical waveguide and recombining the two branch waveguides to form one optical waveguide, and the branch waveguides have an even number of return portions (A1, A2).

The invention discloses an airborne multi-mode cooperative target detection method based on depth measurement, and relates to the technical field of signal processing, and the method comprises the steps: firstly, carrying out the data collection of a target region synchronously through employing an optical camera, a radar sensor and a radiation source sensor which are carried on an airplane, obtaining an optical image, radar data and radiation source data, and carrying out the detection of a target region; carrying out pretreatment on the raw materials; then target detection of respective platforms is carried out on the preprocessed optical image, radar data and radiation source data; and finally, performing alignment and fusion processing on the obtained target detection results of the optical image, the radar data and the radiation source data based on a method of combining depth metric learning with a weighted attention fusion mechanism, and finally obtaining a comprehensive target detection result. The airborne multi-mode cooperative target detection method based on depth measurement provided by the invention not only can improve the accuracy of target detection, but also can effectively improve the robustness and adaptive capacity of the system.

The invention discloses an electromagnetic target feature correction method and device based on adversarial network training, and the method comprises the steps: generating a high-dimensional feature space based on a measurement learning method of center loss and ternary loss, carrying out the characterization and extraction of feature distribution of a data set, and obtaining the data feature distribution of a source domain and a target domain, inputting the target domain data features into a feature correction network for correction, then inputting the source domain data features and the target domain data features into a feature alignment network to estimate a metric distance between the distribution, and performing adversarial training and iterative optimization on the feature correction network and the feature alignment network by taking minimization of the metric distance as a target to obtain a target domain data feature; and reducing the difference between the data feature distribution of the source domain and the target domain until a convergence state is reached. On the premise that the high-dimensional feature space is not changed and the source domain feature distribution is not changed, the distribution difference between the target domain and the source domain features is learned in an adversarial training mode, and correction and learning of the electromagnetic target features under the background environment change condition are completed.

The invention relates to a large-wall-thickness shell forge piece forming method which comprises the following steps that a forming die blank is prefabricated according to the size, the forming die blank is a reducer pipe, the large end of the reducer pipe faces upwards, and the small end of the reducer pipe faces downwards; moulding bed forming; carrying out primary oblique breaking; carrying out secondary inclination breaking; breaking off for the third time; and pressing ring upsetting is conducted, specifically, the pressing ring presses and upsets the tire mold blank till the upper end of the tire mold blank is flush with the upper end of the first tire mold, a forged piece is obtained, the third punch is sleeved with the pressing ring, the pressing ring is a square ring, and the outer diameter of the pressing ring is larger than the inner diameter of an inner cavity of the first tire mold. According to the forming method for the large-wall-thickness shell forge piece, the purposes of material distribution and punching are achieved by prefabricating the forming die blank according to the size and making the blank through the forming die with the first heating number. And the small end of the forge piece is formed through second heating number special-shaped pre-rolling. And forming the large end of the forge piece through third-heating-number tire mold forming. According to the method, the size of the blank is controlled through the die in the whole process, the difference between the manufactured forgings is small, the size state fluctuation is small, the quality consistency of the forgings is good, the heating number is small, operation is easy, and the production efficiency of the shell forgings is improved.

The invention discloses an ICP-OES (Inductively Coupled Plasma Optical Emission Spectrometer) method for determining the content of cerium in aviation gear steel. The method comprises the following steps: setting working parameters of an inductively coupled plasma emission spectrometer; determining the blank solution and the cerium standard solutions under different concentrations by using an inductively coupled plasma emission spectrometer, and drawing a cerium element concentration standard curve; determining a to-be-determined sample solution by using the inductively coupled plasma emission spectrometer, and then determining the content of the cerium element in the to-be-determined solution by combining the cerium element concentration standard curve. According to the method, the content of the cerium element in the aviation gear steel is accurately measured, meanwhile, it is guaranteed that other elements are accurately measured, no element loss is generated, and compared with a steel mill test result, a sulfuric acid-phosphoric acid fuming method, a perchloric acid fuming method and a spectrophotometric method, the difference is small, precision is high, and accuracy and reliability are achieved.

The present invention provides an organic compound, a mixture, a composition, a light extraction layer material and an organic electronic device. The structure of the organic compound is shown in the general formula (1), and the organic compound can be used as a light extraction layer material in an organic electroluminescent device to improve the light extraction efficiency of the device.

This invention discloses a low-light scene analysis method based on multimodal feature fusion and clustering, belonging to artificial intelligence technology. It constructs a single-branch feature extraction network based on Transformer; a multimodal feature interaction and fusion module to achieve multimodal feature interaction and fusion; a multimodal fusion feature clustering module to cluster the features after multimodal interaction and fusion, utilizing the semantic and spatial distances between different features to achieve clustering while simultaneously downsampling the features, mitigating the problem of loss of detailed edge information in structured downsampling; and a multi-scale feature aggregation and decoding module to receive feature information from the encoding network and classify each feature pixel according to the semantic distance of the multi-scale features. This invention can fully utilize visible light and thermal image information and can be applied to scene analysis and navigation of unmanned systems in diverse and complex low-light scenes.

The embodiment of the application provides a kind of training method of image encoder, image processing method, electronic equipment, storage medium and program product, it is related to image processing technical field, the training method of image encoder includes: training image is input image encoder, visual word symbol for representing image feature is extracted by the image encoder;Clothing text word symbol is generated for training image;The clothing text word symbol is used to describe the clothing feature of the person in the training image;The model parameter of the image encoder is iteratively trained using a combination loss function;The combination loss function is used to make the visual word symbol of the same person of multiple training images be smaller, and make the visual word symbol and the clothing text word symbol of the same training image be larger difference.It reduces the degree of dependence on visual information by introducing text modal, increases the precision of image encoder to identify the same person under different clothing, improves the precision and robustness of person re-identification.

The utility model provides an auxiliary tool for adjusting a crystal orientation deviation angle, which comprises a bonding plate for bearing a crystal bar, a base and a mounting plate arranged above the base, the upper surface of the mounting plate is obliquely arranged, and the bonding plate is bonded on the mounting plate. The mounting plate with the inclined upper surface is arranged on the base to be bonded with the bonding plate for bearing the crystal bar, so that the function of adjusting the crystal orientation of the crystal bar in the y direction is realized, the crystal bar with overlarge crystal orientation deviation in the y direction can meet the production precision requirement through the device, and the yield of the crystal bar is improved.

The invention discloses a brightness adjustment method, a display device and a computer readable storage medium, and the method comprises the steps: controlling a display panel to enter a local highlight mode after receiving a target instruction for indicating the display panel to enter the local highlight mode; after a display panel enters a local highlight mode, if a target brightness value of a second area in a display picture is smaller than a preset inflection point brightness value, the brightness value of the second area in the display panel is adjusted to the inflection point brightness value by using a brightness level; adjusting the brightness of a second area in the first picture by using the first calibration value, and adjusting the brightness of the second area in the second picture to a target brightness value by using the second calibration value; wherein the first picture and the second picture are display pictures after the display panel enters the local highlight mode, and the first picture is located in front of the second picture. According to the scheme, the problem that after the display panel enters the local highlight mode, the display brightness is unstable can be solved.

The invention discloses an information transmission method and device, electronic equipment and a readable storage medium, and belongs to the technical field of image processing. The method comprises the following steps: collecting a first video frame; acquiring first information corresponding to the first video frame through a first module of the electronic equipment, wherein the first information comprises at least one of exposure information, automatic focusing information and face detection information; first information is transmitted to a camera application of the electronic device by the first module using a first service for communication between the first module and the camera application.

It is an object of the present disclosure to provide a YFI-type zeolite with a small difference in desorption starting temperature due to hydrothermal durability treatment, a method for producing the YFI-type zeolite, a hydrocarbon adsorbent containing the YFI-type zeolite and a hydrocarbon adsorption method using the YFI-type zeolite The YFI-type zeolite is characterized in that the ratio of the mass of extra-framework aluminum to the mass of contained aluminum is 0% by mass or more and 28% by mass or less.

The invention discloses a transformer and a vehicle, and relates to the technical field of transformers. The second magnetic core is oppositely connected with the first magnetic core; the first coil is wound on the first magnetic core; the coil assembly comprises a second coil and a third coil which are wound on the second magnetic core in parallel, the first coil and the second coil are arranged in a spaced mode to form first leakage inductance, the first coil and the third coil are arranged in a spaced mode to form second leakage inductance, and the first leakage inductance and the second leakage inductance form resonant inductance. According to the technical scheme provided by the invention, the volumes of the transformer and the resonant inductor are reduced.

An electrode apparatus for nerve denervation or modulation includes a main body including a shaft; an electrode unit formed to be drawn out from one end of the shaft and configured to denervate or modulate at least part of nerves on a tube in a body; an electrode guide including a plurality of joint units and a wire connecting the plurality of joint units to each other and configured to guide the electrode unit; and a driving unit located inside the main body and configured to drive the joint units and the wire to protrude from the one end of the shaft. The driving unit drives the joint units in conjunction with the wire to have different displacements.

The invention provides a control method of a railway vehicle, electronic equipment and a storage medium, and relates to the technical field of vehicle control. The railway vehicle comprises a carriage, a hydraulic supporting device and a bogie, the hydraulic supporting device is arranged between the carriage and the bogie, and the method comprises the steps that operation data of the railway vehicle are obtained, and the operation data comprise data reflecting the motion state of the railway vehicle in the running process; according to the operation data, determining whether the carriage rolls laterally; and under the condition that side rolling of the carriage is determined, the state of the hydraulic supporting device is adjusted according to the side rolling direction of the carriage so as to reduce the difference value of the rolling angle between the carriage and the bogie. According to the method, whether the carriage rolls laterally or not can be detected in real time through operation data, and when the carriage rolls laterally, controllable anti-lateral-rolling restoring force can be provided by adjusting the state of the hydraulic supporting device and utilizing hydraulic supporting acting force, so that the carriage tends to be parallel to the bogie, the lateral-rolling amplitude of the carriage is reduced, and high-controllability vehicle lateral-rolling resistance is achieved.

The application provides a gated multi-modal image fusion method and device, electronic equipment and storage medium. The method comprises: acquiring a plurality of modal images corresponding to a target object; inputting each modal image into a feature extraction module of a multi-modal image fusion model for multi-layer feature extraction processing and differential information compensation processing to obtain a plurality of layer feature maps corresponding to each modal image; inputting the third layer feature map corresponding to each modal image into a feature fusion module in the multi-modal image fusion model for feature fusion to obtain an initial fusion feature map; inputting the multi-layer feature map corresponding to the first modal image, the multi-layer feature map corresponding to the second modal image and the initial fusion feature map into a secondary gating fusion module in the multi-modal image fusion model for multi-layer fusion to obtain a target fusion image. Through secondary fusion, the fusion of different modal images can be realized, and the target fusion image information is more balanced.

The present application relates to a kind of facial expression recognition methods based on attention mechanism and self-distillation, comprising the following steps: step 1, constructing facial expression recognition dataset;Step 2, constructing facial expression recognition model, the model is composed of feature extraction module, adaptive channel attention module, remodeling module and self-distillation network;Step 3, the input facial expression image is preprocessed;Step 4, using the training set image in the above facial expression recognition dataset trains facial expression recognition model;Step 5, using the facial expression recognition model trained to the test set and validation set in dataset are inferred classification.The method designs a kind of reasonable and efficient facial expression recognition scheme, it is well guided model to pay attention to those important feature information by the mode of attention mechanism;By a new self-distillation form, refine compressed network knowledge, improve the robustness of shallow network output feature and reduce the complexity of model in inference stage.

Conventionally, a needle or the like needs to be inserted into skin of a subject to invasively collect blood in blood neutral fat measurement, which provides a psychological physical load on a subject. According to the present invention, it is possible to non-invasively estimate the blood neutral fat based on attribute information and / or non-invasive biological information of a predetermined user by generating a blood neutral fat estimation model by machine learning based on attribute information, non-invasive biological information, and blood examination data acquired from a plurality of subjects in advance.

The invention relates to the technical field of auxiliary agents for BC battery manufacturing, in particular to an auxiliary agent for BC battery manufacturing and an application method of the auxiliary agent. The invention relates to an auxiliary agent for manufacturing a BC battery. The auxiliary agent is prepared from the following substances in percentage by mass: 0.3-10% of a high-efficiency dispersant; 0.2%-15% of a surfactant; 0.1%-15% of a stabilizer; 3%-8% of an efficient nucleating agent; 0.1%-5% of a protective agent; and the balance of ultrapure water. According to the auxiliary agent for manufacturing the BC battery, the basic composition of the auxiliary agent for manufacturing the BC battery is optimized, and the protection of the n-poly surface and the p-ply surface of the back surface of the compatible silicon wafer is ensured by adding the protective agent and protecting the n-poly surface and the p-ply surface. After the auxiliary agent for manufacturing the BC battery prepared by the technical scheme of the invention is used, the difference between the textured edge and the middle area of a silicon wafer can be obviously reduced, so that the textured uniformity of the edge wound-plating area and the non-wound-plating area is further improved.

The present application belongs to the technical field of computer vision three-dimensional reconstruction, and specifically relates to a semi-supervised single-view 3D object reconstruction method. A small amount of labeled samples are used to train a neural network, and then the trained neural network is used to generate pseudo labels for unlabeled samples and guide the training of the unlabeled samples. At the same time, the present application proposes that a discriminator scores the quality of the generated pseudo labels, and limits the bias of low-quality pseudo labels on model training. In addition, the present application proposes a prototype shape prior module based on an attention mechanism, which serves as a bridge between images and 3D shapes, reduces the difference between the two modalities, and provides shape prior for the neural network, ensuring that the reconstructed 3D shape is natural. Compared with the current mainstream method in the industry, the present application is more accurate in three-dimensional reconstruction of a single image under the condition that the amount of labeled data is very small, and the generated 3D shape is more realistic and natural.