15results about How to "Simplified Computational Complexity" patented technology

The invention discloses a water supply and drainage automatic control system with intelligent control. Firstly, an acquisition module acquires real-time water source data every a preset time period T; secondly, the intelligent control unit inputs the collected water source data to the trained LSTM neural network model, and the LSTM neural network model performs feature extraction; the extracted water source feature vectors are input into a sparse Gaussian process model, and the sparse Gaussian process model outputs system operation parameters of each water supply and drainage automatic control system in a future time period; and finally, the control execution module adjusts the valve opening degree, the water pump power, the water pressure in the pipeline and the pipeline water flow speed in the corresponding time period according to the system operation parameters output by the sparse Gaussian process model. According to the method, the LSTM neural network is introduced to perform feature extraction of the time sequence data, fine adjustment of water source supply scheduling can be realized in combination with the sparse Gaussian process model, and the user satisfaction is improved.

The invention discloses a deep-sea mining numerical simulation method based on a pseudo-fluid model and related equipment. The method comprises the following steps: firstly, setting an initial velocity field, an initial pressure field and initial particle concentration distribution in the deep-sea mining pipeline; defining boundary conditions of an inlet and an outlet of the deep-sea mining pipeline; finally, numerical simulation is conducted on the deep-sea mining pipeline through the pseudo-fluid model, a deep-sea mining numerical simulation result is obtained, and the deep-sea mining numerical simulation result comprises pressure distribution, a velocity field, particle concentration distribution, effective density and effective viscosity. In the process, the solid particles are simplified into the pseudo fluid by utilizing the pseudo fluid model, so that the effective density and the effective viscosity are introduced, numerical simulation can be performed on the deep-sea mining pipeline by utilizing the effective density and the effective viscosity, and compared with traditional algorithm calculation, the calculation complexity is simplified, and the numerical simulation efficiency is also improved.

The invention belongs to the technical field of image target recognition, and particularly relates to an anti-interference target recognition method combining a pulse neural network and a capsule network, a program, equipment and a storage medium. A capsule structure is introduced into a pulse brain-like model, a self-attention mechanism is introduced, the protocol degree between capsules can be learned in single forward propagation, a capsule attention map is generated through a self-attention layer, and then the capsule attention map is mapped to a classification layer through a one-dimensional convolution layer, so that the model structure and the calculation complexity are simplified, and the accuracy of the model is improved. And irrelevant capsule voting can be selected and inhibited by utilizing a self-attention mechanism, so that information useful for entity recognition is reserved, and the capabilities of long-distance dependence and feature selection are improved. The method has good robustness under the conditions of different noises and different intensities, has good generalization and invariance for affine transformation, and improves the anti-interference capability of target identification.

The invention provides an image splicing method, an image splicing device, a terminal and a computer storage medium, and the method comprises the steps: obtaining all to-be-spliced images, obtaining the actual displacement between two adjacent to-be-spliced images based on a spatial position relation, and obtaining the actual displacement between the two adjacent to-be-spliced images; according to the method, the two to-be-spliced images are subjected to actual displacement, the overlapping area corresponding to the two to-be-spliced images is obtained based on the actual displacement, and the to-be-spliced images are spliced and fused based on the overlapping area to obtain the final spliced image, so that the calculation complexity in the image splicing process is simplified, and the image quality of the spliced image is improved.

The invention provides a knowledge distillation and communication modulation identification method based on a hybrid expert model, belongs to the field of communication, solves the problems of increasing communication modulation identification models and signal interference, and comprises the following steps: training an initial expert model according to a signal-to-noise ratio cutting training set; obtaining a trained hybrid expert model capable of communication modulation identification; training the initial selection model through the training set and the hybrid expert model; if the hybrid expert model is not weighted, identifying an optimal expert model matched with the training set by using T-softmax to obtain a trained selection model, and integrating to obtain a teacher model; if yes, outputting an output result of the teacher model through the aggregation model; and training the initial student model through the output results of the optimal expert model, the training set and the teacher model to obtain a trained student model capable of communication modulation identification. According to the invention, through the trained student model, light weight of the model is realized, and the recognition rate of a low signal-to-noise ratio is improved.

The invention discloses a human organ ultrasonic three-dimensional reconstruction method and device based on respiratory displacement compensation, and the method comprises the steps: controlling a probe at the tail end of a mechanical arm of an ultrasonic robot to collect a plurality of test ultrasonic images in a preset test time period at a target organ of a human body; based on the plurality of test ultrasonic images, carrying out conversion from two-dimensional pixel points to three-dimensional point clouds, and establishing a breathing displacement model of the target organ; and controlling the probe to continuously collect ultrasonic images, and applying the breathing displacement model to all the ultrasonic images to obtain a target organ model after ultrasonic three-dimensional reconstruction. According to the method, the accuracy and the stability of three-dimensional reconstruction are improved, the problem of target organ displacement caused by respiratory movement is effectively solved, the applicability and the reliability of ultrasonic imaging in a dynamic environment are improved, and a more complete and accurate three-dimensional target organ model can be obtained.

The invention discloses a visual inertia filtering method based on manifold processing, and belongs to the field of mobile robot positioning, and the method comprises the steps: obtaining updated pose information based on an inertia IMU sensor; performing Lie group construction based on the updated pose information and the position information of the plurality of landmark points to obtain a complete state of the robot containing deviation information at the current moment; on the basis of the Lie group and the deviation information, uncertain representation, a state model and a measurement model of the current moment of UKF-based invariant Kalman filtering are constructed; updating the state model of the next moment based on the state model; updating a state covariance matrix decomposition factor at the next moment based on the uncertainty of state propagation and the process noise covariance matrix; and if the visual observation data of the landmark exists at the next moment, calculating an observation value based on the measurement model and updating the state and the state covariance matrix decomposition factor at the moment by combining the observation value, otherwise, repeating the step of updating the state model and the state covariance matrix decomposition factor at the next moment.

The application relates to a method and system for realizing 90-degree integer multiple rotation of an image, wherein the method comprises the following steps: an FPGA receives original image data output by an image sensor and image processing parameters issued by an ARM, the image processing parameters at least including width, height and a rotation angle, wherein the rotation angle is an integer multiple of 90 degrees; the original image data is preprocessed and demosaiced to obtain a full-color image in RGB or YUV format; according to the width, height and rotation angle of the image, a coordinate transformation formula for 90-degree integer multiple rotation is used to determine the position mapping relationship between each source pixel in the full-color image and a target pixel after rotation; the full-color image is written into a BRAM cache in a ping-pong mode, and each pixel is read according to the order of the position mapping relationship and written into a target address in a DDR, wherein the target address in the DDR is calculated according to the position relationship.

The application discloses a power electronic distribution network transformer propagation model optimal adaptation method and system, and the method comprises the following steps: calculating transformer winding leakage inductance and winding resistance, and transformer excitation inductance and excitation resistance; estimating the equivalent value of each link distributed capacitance; deducing the propagation coefficient of the high-frequency transformer model and the low-frequency transformer model at any frequency under the excitation source; calculating the relative deviation amount of the corresponding propagation coefficient of the high-frequency transformer model and the low-frequency transformer model at different frequencies in the required frequency band of the simulation scene; obtaining the maximum value of the relative deviation amount of the propagation coefficient under all conditions in the required frequency band of the simulation scene; when all the maximum values of the relative deviation amount of the propagation coefficient are less than the set relative deviation limit value, the low-frequency transformer model is selected for the simulation scene, otherwise the high-frequency transformer model is selected; the application has the advantages that the optimal selection of the transformer model is performed under the consideration of the simulation efficiency and accuracy requirements.

The embodiment of the invention provides an index management display method and system and electronic equipment, and relates to the technical field of index management. The index management display method comprises the following steps: setting a plurality of pages at the front end of VUE; an index correlation table is stored in the JAVA rear end; the index correlation table comprises an index configuration table, an index detailed configuration table and an index landing library table; acquiring operations of a user at the front end of the VUE based on the plurality of pages, wherein the operations comprise a configuration operation, a calculation operation and a query operation; mapping the configuration operation into a table operation for the index correlation table, mapping the calculation operation into a calling execution operation for an index calculation program, and forwarding the query operation to a corresponding interface of a data warehouse where the index correlation table is located; and obtaining an execution result, and feeding back the execution result to the front end of the VUE as feedback of the operation. According to the embodiment provided by the invention, the problem of tedious index configuration of a service user is solved.

The application discloses an image feature extraction method and an image noise reduction method and related devices, and belongs to the technical field of image processing. The application can strengthen the connection between the local area and the global information of the features from the progressive angle of the large local receptive field and the global receptive field, so that the feature extraction method based on the self-attention mechanism can still capture the long-distance pixel dependence while reducing the calculation complexity, can well remove the image noise in the image noise reduction task based on the pixel level, and can maximize the retention of the detail information of the image.

The present invention discloses a dynamic fuzzy neural network modeling method for intelligent dimming of an LED system, which relates to the technical field of nonlinear optical characteristics of LED systems, and includes the following steps: pre-calibrating the input variable characteristics and expected output value characteristics of the photoelectric thermal model of an LED module; building and training a dynamic fuzzy neural network model, wherein the dynamic fuzzy neural network model includes an input layer, a fuzzification layer, a precondition layer, a postcondition layer, and an output layer, and the training of the dynamic fuzzy neural network model includes network structure optimization and parameter learning. The present invention adopts an online modeling method, and the model structure parameters can be optimized and adjusted in real time, which has stronger adaptability. The combined measurement of heat sink temperature and ambient temperature is used to indirectly reflect the influence of junction temperature on the photometric output of the LED system, thereby avoiding the complexity of junction temperature measurement.

The invention discloses a power distribution network load transfer auxiliary decision-making method based on a partition autonomous model, and the method comprises the steps: carrying out the modeling of a power grid simulation example through offline power grid data, and carrying out the partitioning of the simulation example, and obtaining an autonomous region range; constructing an autonomous model according to the offline power grid partition data, carrying out model training on the offline fault data, and obtaining a converged auxiliary decision-making model; obtaining online power grid data, and preprocessing the power grid data to obtain an online power grid data partitioning result; and according to the actual fault scene of the power grid, calling the trained model to carry out load transfer online auxiliary decision making. A power grid partition autonomy and interval cooperation method is adopted, partition standards are constructed through multiple pieces of power grid information such as electrical distance and inter-bus voltage sensitivity, cooperation reward functions of different autonomous region models are constructed according to power grid operation specifications, and a preprocessing method of power grid online data is designed based on data specifications of an offline scene. And finally, the auxiliary decision-making of the load transfer after the power grid fault is realized.

The invention discloses a temperature adjusting module, electronic equipment and an error compensation method, and belongs to the technical field of sensors, and the temperature adjusting module comprises a first sensor, a first temperature detection device, a temperature compensation device and a heat source; a heat propagation path is arranged between the heat source and the first sensor; the first temperature detection device is arranged corresponding to the first sensor and is used for detecting the temperature of the first sensor; the temperature compensation device is electrically connected with the first temperature detection device and is used for performing temperature compensation on the first sensor according to the temperature detected by the first temperature detection device so as to adjust the temperature change rate of the first sensor to be consistent with a target temperature curve; wherein the target temperature curve is used for determining a temperature hysteresis curve of the first sensor, and the temperature hysteresis curve is used for compensating offset errors caused by temperature changes of the first sensor.