31results about How to "Effective decoupling" patented technology

A hydrogen energy storage power station control system adopts a layered structure, comprehensively judges from electricity price information, renewable energy output, environment temperature and hydrogen storage tank state data through a reinforcement learning model, selects a hydrogen production, power generation or standby mode, and sets power targets of an electrolytic cell and a fuel cell; the middle layer uses an optimization algorithm to refine power distribution, the lower layer executes a power instruction through an industrial control algorithm, and an independent safety guarantee layer is set to monitor indexes such as hydrogen concentration, pressure and temperature. In the optimization process of the system, energy consumption and equipment life loss of an electrolytic bath and auxiliary equipment are considered at the same time, full-life-cycle economy is achieved, the economy of a hydrogen energy storage power station is improved, and the service life of key equipment is prolonged.

The invention relates to a high-precision steel surface defect detection method suitable for a complex industrial environment, and aims to solve the problems of feature coupling and background interference caused by data scarcity, long tail distribution and multi-defect coexistence. According to the method, a conditional generative adversarial network (CGAN) and a convolutional neural network (CNN) are combined, an SE-Net channel attention mechanism is integrated, and the method is used for automatic detection and classification of multiple defects on the steel surface. By introducing a physical constraint CGAN data generation method, a multi-defect coexistence composite image conforming to industrial reality can be generated, so that the diversity and accuracy of a training data set are effectively enhanced. The model optimizes the expression of defect features through multi-level feature extraction and an attention mechanism, and enhances the distinguishing ability and recognition precision of multiple defects. Experiments show that the method is excellent in performance in a multi-defect identification task, the accuracy rate of 98.89% and the F1 score of 99.72% are achieved, the method is remarkably superior to a traditional detection method and other deep learning models, and the method has high robustness and practical value in a complex industrial environment.

The invention discloses a small-scale deformation InSAR time sequence inversion method and system based on multiple stable points, and the method comprises the following steps: S1, carrying out the small baseline interference processing of multi-temporal synthetic aperture radar data, and screening high-coherence pixel points, S2, recognizing a stable point candidate set with dense spatial distribution; s3, constructing an initial atmospheric delay phase model covering a full interferogram range; s4, obtaining an interferometric phase after atmospheric correction; and S5, inverting the deformation rate and the deformation time sequence parameters of the InSAR again. According to the method, dense stable points are screened based on a deformation rate threshold value, and a space atmosphere model is constructed, so that the problem that atmosphere delay and nonlinear deformation are difficult to separate in InSAR time sequence analysis is effectively solved, and high-signal-to-noise-ratio and high-precision time sequence monitoring of small-scale weak deformation of the earth surface is realized.

This invention provides a fault detection method and system based on low-rank decomposition and common trend decoupling. The method first acquires historical process data under normal operating conditions of an industrial process as training data. Then, it decomposes the training data into clean data representing the main structural information and sparse noise data representing noise interference using low-rank decomposition. Next, it constructs a common trend learning model based on the clean data, reconstructs the non-stationary common trend, and uses the residual between the clean data and the non-stationary common trend as a stationary feature. Decoupling is achieved by applying stationarity constraints. Further, a fault detection model is obtained through joint optimization. In the detection stage, the process data to be detected is input into the model, and monitoring statistics based on the reconstruction error of the stationary feature and the common trend are constructed respectively. The result of the control limit comparison determines whether a fault has occurred, thereby improving the robustness of fault detection in non-stationary industrial processes.

The application provides a security-constrained economic dispatch method based on a dual Halpern iteration, and belongs to the technical field of power system optimization dispatch and artificial intelligence computing infrastructure. The method comprises the following steps: constructing an original model of a security-constrained economic dispatch problem SCED; modeling the renewable energy output uncertainty by using a Gaussian mixture model, converting the chance constraint in the original model into a deterministic linear constraint, and thus converting the original model into a convex quadratic programming problem; constructing a limited Wolfe dual model for the convex quadratic programming problem, solving the dual model by using a Halpern iteration, and obtaining a dispatch optimization result. The application can solve the problems of low calculation efficiency, large memory overhead, low GPU utilization rate and difficulty in stable solution in a super-large-scale scenario in large-scale SCED, and can quickly generate an executable dispatch scheme, so that the calculation efficiency and engineering applicability of online dispatch and rolling optimization are improved.

The application relates to the technical field of electronic component reliability, and discloses a flat transformer voltage withstand performance sure reliability modeling and evaluation method, which comprises the following steps: firstly, a constitutive model fusing a physical mechanism is constructed, a time degradation model and a temperature influence model are established based on the geometric size, material properties and environmental stress of an insulating layer, and a process capability parameter is introduced to represent manufacturing fluctuation; secondly, a reliability evaluation model is established, a failure threshold and a margin equation are defined, and physical parameters are identified as random variables to quantify uncertainty; then, Monte Carlo algorithm is used to sample random variables, virtual samples are generated, and voltage withstand margin is calculated on an evaluation time sequence; finally, the number of samples meeting a safety criterion is counted, instantaneous reliability is calculated, and a full-life-cycle reliability curve is generated. The application effectively decouples and quantifies the influence of manufacturing processes, temperature stress and time degradation on voltage withstand performance, and realizes sure reliability evaluation with high confidence.

This invention belongs to the field of vehicle control technology and relates to a vehicle motion control system based on angular module dynamics. The system constructs a two-degree-of-freedom driver control system including steering wheel input and vehicle posture lever input, used to represent path curvature tracking requirements and vehicle posture adjustment requirements, respectively. Based on the driver control input, ideal yaw rate control targets and ideal center-of-gravity sideslip angle control targets are constructed, and the four-wheel steering angle feedforward control quantities are determined through vehicle kinematics methods of pure yaw motion and pure lateral motion. Through feedback optimization control, the additional yaw moment and four-wheel steering angle control quantities are coordinated under composite operating conditions. Based on adaptive adjustment of control weights and constraints under operating conditions, the distribution control of four-wheel drive / braking torque and four-wheel steering angle is achieved. This invention achieves decoupled control of vehicle yaw motion and lateral motion by constructing a two-degree-of-freedom driver control system and combining feedforward kinematics and feedback dynamics control.

The invention relates to the technical field of data analysis, in particular to a palmprint recognition anti-fraud method based on double-branch self-supervised learning, and the method comprises the steps: constructing a double-branch self-supervised reconstruction architecture, determining a first to-be-processed image based on a high-frequency self-supervised branch, and determining a second to-be-processed image based on a chromaticity regularization self-supervised branch. Dividing the two images into patch blocks which are not overlapped with each other, independently generating random mask matrixes with different spatial distributions, determining visible areas, inputting the patch blocks of the visible areas into a shared encoder for feature extraction, outputting latent variable features, and splicing the latent variable features with learnable mask marks to obtain a patch matrix; and respectively sending to a high-frequency decoder and a chroma decoder, determining total loss, carrying out physical decoupling on illumination and material attributes, and carrying out fine adjustment on the architecture to finish convergence of the architecture. According to the method, the palm print features with discrimination are extracted through the shared encoder, and chromaticity distribution consistency constraints are introduced, so that effective decoupling of ambient light and real material attributes is realized.

This invention discloses an underwater wireless optical communication alignment method and system based on spatial coherence adaptive differential observation. The method includes: constructing a photodetector array in the receiver, constructing a normalized differential observation structure, and through differential observation and dynamic spacing adjustment scheme, a laser system transceiver alignment method that can effectively suppress turbulent multiplicative scintillation and adapt to different water dynamic conditions can be developed, significantly improving the alignment accuracy and robustness between the receiver and the transmitter in complex turbulent environments.

The application relates to an air-ground cooperative navigation control method and system based on global priori of an unmanned aerial vehicle and elastic preset performance tracking of an unmanned vehicle, and belongs to the technical field of autonomous navigation and cooperative control of unmanned systems. In view of the problems of a single system, such as a large blind area of sensing, a prominent calculation bottleneck, easy falling into a local minimum value in local obstacle avoidance, weak anti-interference performance in trajectory tracking and waste of communication resources in a complex environment, a layered cooperative architecture is adopted, the unmanned aerial vehicle is used to complete mapping, two-dimensional projection and global path planning and issue a discrete topological path, the unmanned vehicle generates a local trajectory through spatial domain robust differential game, dynamic optimization is realized by combining elastic preset performance control and internal and external double-loop reinforcement learning, high-precision tracking and global optimization of energy consumption are realized, and an event triggering mechanism is used to reduce communication load. The application improves navigation reliability, trajectory stability and resource utilization efficiency in a complex dynamic environment.

This invention discloses a method for measuring Rydberg atoms using micro-Doppler agility under a polarization-multiplexed microwave field, belonging to the field of quantum precision measurement technology. The method includes generating a dynamic coded sequence and an interleaved coded microwave field by aperiodically pairing and jumping the microwave polarization state with the carrier frequency; acquiring the target echo signal and applying it to the Rydberg atom ensemble with a reference microwave field to generate a quantum coherent state; applying a synchronous laser pulse to acquire a time-series optical response signal; obtaining the signal's fundamental frequency, envelope spectrum, and initial phase component through time-frequency transformation and feature separation; and performing parameter mapping and synthesis to generate three-dimensional micro-motion parameters. This invention uses an aperiodically agile polarization-frequency-multiplexed microwave field for active target detection and utilizes the Rydberg atom ensemble for quantum coherent detection and optical readout of the echo signal, enabling high-sensitivity, high-precision, and high-anti-interference measurement of the target's three-dimensional micro-motion parameters.

The invention discloses an anthropomorphic robot multi-part expression cooperative control method, device, system, equipment and product, and relates to the technical field of anthropomorphic humans. The method comprises the steps of firstly obtaining a target expression control signal, then responding to the signal to obtain one-to-one corresponding and mutually independent silence delay parameters for at least two expression execution mechanisms, then synchronously starting a plurality of silence delays based on the parameters, and finally respectively driving the corresponding execution mechanisms when the delays are finished. Therefore, by introducing independent and adjustable silent delay parameters and synchronously starting, the problems of'programmed sequence sense 'and'mechanical cooperation' caused by asynchronous distribution of multi-part action starting instructions in the prior art are solved, and the biological characteristics of'synchronous response and asynchronous execution 'of human facial neural signals are effectively simulated; therefore, the naturalness and life feeling of robot expressions are remarkably improved, and practical application and popularization are facilitated.

This invention discloses a method, device, and medium for querying the capacity of logical units based on pre-allocation, relating to the field of distributed storage technology. By pre-allocating physical storage blocks and establishing mapping relationships for logical storage units that are not bound to physical storage blocks when responding to capacity query commands, the system can accurately calculate the available capacity based on the actual physical attributes of the physical storage block, solving the problem of inaccurate queries. Simultaneously, a pre-allocation state is introduced, distinguishing this state from the conventional allocated or active states. Physical storage blocks in the pre-allocated state are considered idle in the resource reclamation statistics of the storage device. When a logical storage unit receives a data write command, the pre-allocation state is changed back to the allocated state, ensuring the correctness of physical block lifecycle management. While ensuring high accuracy in capacity queries, efficient resource scheduling is achieved, improving the overall reliability and performance stability of the storage system.

The present application relates to the technical field of predictive maintenance and fault diagnosis of industrial equipment, and particularly relates to a multi-motor coupled vibration intelligent diagnosis method and system. The technical scheme comprises the following steps: based on the collected vibration data of the multi-motor under the normal operation state and the physical parameters of the motor equipment, a parameterized fault simulation module is used to generate a coupled vibration simulation data set with accurate fault labels. The present application realizes the cold start of the diagnosis system under the condition of no historical fault samples; the deep reinforcement learning agent is used to process the space-time coupled signal, and the layered reward mechanism is combined to significantly improve the decoupling diagnosis accuracy and early warning capability; through the online adaptive learning module, the system can be fine-tuned in real time according to the field feedback, and quickly adapt to the individual equipment state; the simulation-model evolution module uses the accumulated real data to continuously optimize the simulator and the diagnosis model, and forms a self-iterating closed loop.

The present application belongs to the technical field of image processing, and particularly relates to an artistic style transfer method and system based on frequency domain modulation and attention guidance, aiming at the difficulty of effectively separating content structure and style features in the prior art, which easily leads to structural damage or insufficient stylization of the generated results. The present application first initializes the strategy based on structure perception frequency domain modulation, effectively decouples content and style from the frequency domain; constructs a decoupling cross-attention mechanism, processes text, content and style features through parallel channels respectively, thereby avoiding feature interference while precisely regulating the generation process; and uses an attention-driven guidance mechanism to model the style transfer as an iterative optimization problem in the latent space, dynamically calibrates the feature distribution through gradient guidance, and effectively alleviates the feature dilution problem in the deep network.

The composite film is formed by laminating and compounding at least two layers of polymer porous films with anisotropic microporous structures, and the adjacent film layers are stacked after rotating by 45-135 degrees around the axis perpendicular to the film surface, the micropore channels of the upper and lower film layers form a non-through structure in a three-dimensional space, and decoupling of a heat flow transmission direction and a gas diffusion direction is realized. The interlayer dislocation degree is controlled by regulating and controlling the rotation angle, an efficient thermal resistance barrier is formed in the thickness direction, and meanwhile a gas diffusion channel is kept in the plane direction. The composite film has excellent heat insulation performance and air permeability, the heat conductivity coefficient can be as low as 0.028 W / (m.K), the water vapor transmission rate is kept at 980 g / (m.24 h) or above, the preparation process is simple and controllable, and the composite film can be widely applied to the fields of automobile interiors, building heat preservation, garment materials and the like.

The application relates to the technical field of intelligent control of underwater robots, and provides a bionic underwater robot motion control method and system based on asymmetric information. The bionic underwater robot motion control method based on asymmetric information comprises the following steps: first, an underwater robot is abstracted as a double-control system with two sub-controllers of a wave fin and a tail propulsion mechanism. In view of the characteristics that the motion state of the underwater robot cannot be directly measured and the environment is unstructured, the output of acoustic, inertial and other sensors is modeled as a multi-modal linear perception model. Based on the model, corresponding asymmetric measurement signal fusers are designed for each controller. In order to solve the "fusion-control" coupling problem caused by asymmetric information, the random maximum value principle is adopted to realize the decoupling design of the fusers and the controllers, and finally an explicit solution of the optimal decentralized controller is given. The control precision and robustness of the system in a dynamically uncertain environment are significantly improved.

The invention provides a sensitive attribute forgetting method and device for a recommendation system. An embedding layer of the recommendation system comprises a user embedding matrix and an article embedding matrix; the method comprises the following steps: performing matrix decomposition on an article embedding matrix to obtain an independent space orthogonal basis; performing feature decoupling on the user representation vector to obtain a correlation vector and an irrelevant vector; the irrelevant vector is represented as a linear combination of a to-be-optimized parameter and an irrelevant space orthogonal basis; constructing user representation distribution and sensitive attribute distribution, and selecting the optimal bandwidth of the HSIC core; constructing a forgetting loss function used for measuring the dependency between the user representation distribution and the sensitive attribute distribution, and replacing a user representation vector in the forgetting loss function with a combination of a correlation vector and an irrelevant vector to obtain a target function; and under the condition of keeping the correlation vector unchanged, taking a minimization objective function as an objective, obtaining a final user representation vector through iterative optimization, and taking the final user representation vector as a result after the sensitive attribute is forgotten.

The application discloses a quail egg intelligent detection method and system based on multi-source data fusion, and relates to the technical field of quail egg quality detection. The quail egg intelligent detection method based on multi-source data fusion extracts image texture features and spectral response features by acquiring visible light images, near-infrared images and geometric parameter data of quail eggs, analyzes image equalization indexes and structure transmission indexes respectively, analyzes geometric morphological indexes in combination with morphological parameters, and constructs comprehensive state indexes, thereby determining the quality grade of the quail eggs to be detected. The application extracts image equalization indexes, structure transmission indexes and geometric morphological indexes by fusing visible light images, near-infrared images and geometric parameters, constructs comprehensive state indexes, can comprehensively evaluate the quality of quail eggs from multiple dimensions such as appearance, light transmission and morphology, effectively identifies internal abnormal conditions that are prone to be missed in traditional image detection, and improves the accuracy of quality grade judgment and the stability of overall detection results.

This application relates to an AI model image generation method, system, device, and medium, belonging to the technical field of image processing. It involves collecting multimodal data of the target garment, including mannequin images, flat static images, and material close-up images; selecting a reference image from a target pose library whose pose matches the desired pose, and extracting the corresponding pose skeleton image as the target pose skeleton image; inputting the flat static image and the material close-up image into a visual language model to generate a text description; and inputting the target pose skeleton image, mannequin image, and text description into a pre-trained LoRA diffusion model to generate a preliminary dressed image. The LoRA diffusion model's input layer includes an independent pose condition injection channel, through which the target pose skeleton image is input into the LoRA diffusion model. This application has the beneficial technical effect of avoiding stiff and unnatural AI model poses, thereby improving the aesthetics of the generated images.

This invention provides a radio frequency fingerprinting method, device, and medium based on a complex neural network. The radio frequency fingerprinting method based on a complex neural network includes: converting acquired raw complex channel state information (CSI) data into amplitude data, and adding all-zero imaginary parts to the amplitude data to construct complex form data conforming to the input format of a complex neural network; dividing the complex form data into a training set and a test set; constructing a complex neural network model for feature extraction from the input complex form data; training the complex neural network model using the training set; inputting the CSI data to be identified, after being constructed into complex form data, into the trained complex neural network model, and determining the device category to which the CSI data to be identified belongs based on the amplitude value of the complex result output by the model. This achieves accurate and stable identification of device identity in complex wireless environments.

The application discloses an adversarial knowledge distillation method and system based on feature semantic guidance and feature orthogonalization, which firstly acquires an image classification dataset, extracts text features, and constructs a semantic feature space.Secondly, a knowledge distillation framework including a teacher model, a student model and a projection head is constructed, the input of the teacher model and the student model is an image in the dataset, and the output is an image category prediction result. Then, the student model parameters are frozen, the projection head is updated through momentum tracking, in the feature level semantic adversarial sample generation stage, all network parameters are frozen, and normalized disturbance is constructed to obtain adversarial features. Finally, in the student model multi-objective joint optimization stage, the teacher model and the two projection heads are fixed, the student model parameters are updated, the classification loss, the distillation loss, the consistency loss and the orthogonalization loss are introduced, and optimization is completed. The application improves the fine-grained discrimination and generalization ability of the student model image classification, and does not introduce additional reasoning overhead.

A SiC MOSFET gate oxide layer degradation comprehensive diagnosis method and system, the method comprises the following steps: constructing a data set, and training to obtain a junction temperature estimation function; obtaining actual precursor parameters under actual operating conditions through a measurement circuit, substituting the actual precursor parameters into the junction temperature estimation function to obtain a current actual junction temperature estimation value, substituting the current actual junction temperature estimation value into a linear relationship between the precursor parameters and the junction temperature before the gate oxide layer degrades to obtain the flat band voltage and the threshold voltage of the gate oxide layer without degradation at the junction temperature; calculating the difference between the actual precursor parameters and the precursor parameters at the junction temperature without degradation of the gate oxide layer to obtain a precursor parameter drift amount, and performing comprehensive diagnosis of the gate oxide layer degradation according to the precursor parameter drift amount to obtain a comprehensive diagnosis result. The present application solves the misdiagnosis problem caused by temperature drift in the prior art, improves the accuracy and robustness of the degradation diagnosis, and realizes real-time monitoring and thermal state management of the device working junction temperature state.

The invention discloses an RTP equipment temperature control method and device based on diagonal matrix decoupling, and relates to the technical field of semiconductor manufacturing temperature control. Aiming at the problems of strong coupling and large time delay of multiple temperature zones of RTP equipment and difficulty in accurately solving a model inverse matrix, the method comprises the following steps: firstly, establishing a multi-temperature zone heating model; solving an approximate inverse matrix of the heating model by using a relative normalized gain matrix (RNGA) and a relative average residence time matrix (RARTA); a time delay correction strategy is introduced, a diagonal matrix decoupling model is constructed, and physical realizability is ensured; and finally, closed-loop control is realized in combination with an incremental PID controller. According to the method, the problem that the decoupling matrix of a high-dimensional coupling system is difficult to calculate can be effectively solved, overshoot is remarkably inhibited, and the wafer temperature uniformity and the anti-interference capability are improved.

The embodiment of the invention relates to an electrostatic sensor particle coupling signal deconvolution integral peak method. The method comprises the following steps: S1, obtaining an electrostatic detection signal fragment containing a multi-abrasive-particle overlapping feature; s2, constructing a Gaussian mixture model of the electrostatic detection signal segment, wherein the Gaussian mixture model is formed by superposing a plurality of single Gaussian components; s3, estimating an initial parameter set of the Gaussian mixture model based on waveform geometric features of the electrostatic detection signal segments; s4, performing global search on the initial parameter set by using a global optimization algorithm to obtain a parameter approximate solution, taking the parameter approximate solution as an initial value, performing optimization by using a local optimization algorithm until a preset convergence condition is met, and obtaining an optimal parameter set; and S5, reconstructing each independent single Gaussian component according to the optimal parameter set, and determining the feature information of the single abrasive particle according to each independent single Gaussian component.

This invention relates to the field of power equipment monitoring and fault diagnosis technology, and discloses a method and system for early warning of power transformer faults. The method includes collecting the transformer's load current, ambient temperature, and top oil temperature sequences; calculating the initial measured temperature rise based on the top oil temperature and ambient temperature, and determining the thermal response lag time through cross-correlation analysis; shifting and correcting the load current according to the lag time, and constructing a time-aligned excitation input set and measured temperature rise sequence based on a common effective time window; substituting the excitation input set into a thermal circuit model based on the thermal balance differential equation to calculate the theoretical reference temperature rise, combining the measured temperature rise sequence to calculate the residual and performing moving average filtering to construct a fault feature vector; inputting the fault feature vector into a fault diagnosis classification model to determine the fault type and severity, and generating a graded early warning command accordingly. This method can eliminate thermal lag interference and achieve accurate quantitative early warning of latent thermal faults in transformers under dynamic operating conditions.

The application relates to a three-axis gyroscope, in particular to a novel three-axis silicon micro gyroscope. The application solves the problems of low measuring precision and high production cost of the existing three-axis gyroscope. The novel three-axis silicon micro gyroscope comprises a glass substrate, a resonator part and an electrode part; the resonator part comprises a cylindrical center anchor point, four groups of square anchor points A and four square anchor points B; the cylindrical center anchor point, the four groups of square anchor points A and the four square anchor points B are all bonded to the upper surface of the glass substrate; the electrode part comprises four pairs of arc-shaped electrodes A, four arc-shaped electrodes B, four pairs of arc-shaped electrodes C, four arc-shaped electrodes D and four rectangular electrodes; the four pairs of arc-shaped electrodes A, the four arc-shaped electrodes B, the four pairs of arc-shaped electrodes C and the four arc-shaped electrodes D are all bonded to the upper surface of the glass substrate; and the four rectangular electrodes are all sputtered on the upper surface of the glass substrate. The application is suitable for high-precision and high-tech fields such as military navigation and deep space exploration.

This invention relates to a UAV target recognition method based on ISAC signal decomposition and background contrast attention fusion, belonging to the field of radar signal processing. The method includes: configuring the transmitted signal of an ISAC base station based on a 5G NR frame structure; receiving and establishing a CSI mixed signal model; constructing a parameterized operator; using the rmD-NSP algorithm to iteratively decompose the mixed signal, separating the rotor micro-Doppler component, and converting the separated signal into a micro-Doppler time-frequency spectrum; using the RMBL module to model multi-scale background features of the time-frequency spectrum; using the BCA module to calculate the difference weights between target features and background features, generating a background-suppressed micro-motion feature vector; generating an HRRP sequence using the CSI mixed signal; using an LSTM network to extract the target's geometric structure feature vector; and weighted concatenating and fusing the micro-motion feature vector and the geometric structure feature vector and inputting them into a classifier to output the UAV target's category probability.

The invention discloses a wind power converter open-circuit fault identification method, and the method comprises the steps: constructing a dynamic health state baseline in real time through employing q-axis and d-axis current instructions capable of accurately reflecting the dynamic intention of a system, and carrying out the feed-forward compensation of Park vector module length, thereby actively stripping the interference of the change of a normal working condition on fault features, and improving the fault identification precision. Therefore, the compensated fault features can be kept stable in a healthy state, obvious deviation is generated only when a real fault occurs, effective decoupling of the fault features and system transient response is achieved, the signal-to-noise ratio of diagnosis is obviously improved, and safe and stable operation of a wind power generation system is effectively guaranteed.

The application provides a 4D video generation method, device and equipment with controllable camera and illumination, relates to the technical field of computer vision and computer graphics, and aims to solve the problem that the existing model cannot simultaneously control the camera trajectory and the illumination condition. The method comprises the following steps: based on an input video, generating a dynamic point cloud and a sparse heavy-light point cloud, the dynamic point cloud is used for explicitly representing the geometric structure and motion information of the scene in the input video, and the sparse heavy-light point cloud is used for providing the illumination prior of the scene under the target illumination condition; according to a target camera trajectory, the dynamic point cloud and the sparse heavy-light point cloud are processed respectively to generate geometric prior information and illumination prior information aligned with a target view angle; the geometric prior information and the illumination prior information are encoded into visual tokens, and illumination query tokens are extracted from heavy-light key frames; denoising generation is performed by using the visual tokens and the illumination query tokens, and a target video consistent with the target camera trajectory and the target illumination condition is output.