21results about How to "Reduce computing requirements" patented technology

The invention discloses an advanced data encryption protection technology for DNA information storage, and the specific technical scheme of the technology comprises the steps: coding to-be-encrypted data into small information blocks and transcoding the small information blocks into a DNA sequence through a coding and encryption scheme, after the DNA sequence is synthesized, introducing interference information through random errors or fractures, the interference intensity of random combination interference noise is effectively improved; the method for decrypting the DNA library carrying the encrypted information comprises the steps of sequencing, fragment assembly, information small block reconstruction, reordering and decoding, and finally decoding the original information based on an erasure coding scheme. The technology has a wide application prospect in the field of data storage and protection.

The invention discloses a ship route planning method and system based on deep learning large-scale extreme wind speed forecast, and relates to the technical field of artificial intelligence and maritime navigation. An incremental principal component analysis method is introduced, a large-scale sea surface extreme wind speed forecasting model based on depth separable convolution is constructed, and potential physical sudden change disturbance information is decoded and reserved in a variable high-dimensional mapping space by the model through integration of an attention mechanism module and a residual learning algorithm; the attention mechanism module maps high-dimensional feature information in an extreme wind field time sequence to a multi-dimensional feature matrix of a coder-decoder. The method can perform refined dynamic forecasting on the large-scale sea surface extreme wind speed, accurately detect the extreme wind speed sea area and the boundary thereof, forecast and master the intermittent sudden change rule of the sea extreme wind speed in advance, help a ship driver to dynamically adjust the ship route in real time and avoid severe storm weather, and further ensure the navigation safety.

The present invention discloses a digital multi-beam forming method based on multi-domain combination. The multi-domain combined digital multi-beam technology is adopted, and the space-time two-dimensional domain is used for joint anti-interference. A constraint vector pointing to the satellite is formed according to the satellite direction information and array flow, so that the beam formed after anti-interference is aligned with the satellite direction, and a null is formed in the interference direction while protection can be formed in the useful signal direction, so that the signal processing gain brought by the antenna array can be utilized while the interference is suppressed, thereby improving the signal-to-noise ratio of the satellite signal. The broadband correction technology is used for time domain compensation, and the in-band amplitude and phase inconsistency brought by the hardware is corrected, thereby avoiding the influence of the carrier phase difference between different beams brought by the time domain of the space-time multi-beam technology. While improving the anti-interference performance, the positioning accuracy of the multi-beam signal is guaranteed, and the requirement for the hardware in-band consistency is reduced, so that the space-time multi-beam algorithm is simple and feasible in engineering applications.

The invention discloses a fault diagnosis method and system for a motor driving mechanism of a high-voltage circuit breaker, relates to the technical field of high-voltage switch equipment, and comprises an intelligent fault diagnosis method combining interactive multi-model and unscented Kalman filtering. Unscented Kalman filtering is utilized, so that the nonlinear characteristic of a state equation under the fault condition can be better approximated, and the filter can have better stability and lower calculated amount requirements; by means of the interactive multi-model, the problem that fault diagnosis is difficult due to the fact that measurable parameters are few is solved, and the rapidity and accuracy of fault diagnosis under the condition that the difference between the generated fault and the pre-assumed fault is large are improved. In the diagnosis process, an active matching mechanism is used for improving the problem that the matching degree of a fault model and an actual fault is poor, faults are actively found or eliminated, and the accuracy and rapidity of fault diagnosis are improved.

The invention provides a broadcast signal broadband sorting method, which comprises the following steps of: S1, collecting signal spectrum data in a preset filtering range for multiple times, and setting a self-adaptive threshold; s2, de-noising the signals according to the self-adaptive threshold, and dividing the de-noised signals into signals containing carrier waves and signals not containing carrier waves; s3, analog broadcast signal sorting is carried out on the signals containing the carriers, and analog broadcast signals are output; and S4, carrying out digital amplitude modulation broadcast signal sorting on the signal without the carrier, and outputting a digital broadcast signal. According to actual electromagnetic environment conditions, broadcast signal features are combined, the broadcast signal sorting process is simplified, and the broadcast signal sorting rate is improved under the condition that the recognition rate is not reduced.

The invention relates to the technical field of radar testing, and discloses a weather radar speed spectrum width testing system and method.The weather radar speed spectrum width testing system comprises a weather radar, detection devices and a ground console, the detection devices comprise the first detection device and the second detection device, and the first detection device comprises a first movement mechanism, a first fixing mechanism, a first measuring mechanism and a releasing mechanism; the second detection device comprises a second movement mechanism, a second fixing mechanism and a second measurement mechanism, the ground control console is provided with a wireless control module, a data receiving module and a data comparison module, and a first test is completed based on the weather radar, the first detection device and the ground control console; and completing a second test based on the weather radar, the first detection device, the second detection device and the ground console. The first test is completed based on the first detection device, and the second test is completed based on the first detection device and the second detection device, so that the radar test is more complete, and the accuracy is higher.

The application discloses an interface processing method and device, which are applied to an interface processing scene. The method comprises the following steps: in response to an interface recording request of a user end, recording data corresponding to the interface recording request is acquired; the recording data is sent to a management platform end, so that the management platform end generates an interface set according to the recording data; an interface request is generated, and the interface request is sent to a proxy server, so that the proxy server calls the interface set generated by the management platform end to obtain target simulation data responding to the interface request; and the target simulation data sent by the proxy server is received. Through the automatic and centralized interface recording and simulation mechanism, the efficiency and accuracy of the business link simulation are improved, and the maintenance cost is reduced and the flexibility is improved.

The present application relates to absorption spectroscopy technology, and specifically provides a multi-gas measurement method based on GFC wheel, comprising the following steps: A1. obtaining standard deviation R of reference signal in multiple detection periods rsd and standard deviation M of measurement signal rsd ; A2. obtaining offset parameter K=f(R rsd , M rsd , t) according to the standard deviations, t is the absolute value of the difference between the rotation period at the last calibration and the current rotation period; A3. comparing the offset parameter K with a set range; if the offset parameter K is normal, no adjustment is needed; if the offset parameter K is abnormal, corresponding adjustment measures are implemented according to the set range where the offset parameter K is located, and the adjustment measures comprise: obtaining complete reference signal and measurement signal, thereby obtaining Δt, and replacing. The present application has the advantages of accurate measurement and the like.

The invention provides a cross-space collaborative learning lightweight coast dam crack segmentation method, and the method comprises the steps: S1, carrying out the down-sampling through a three-layer encoder, unifying the channel number and resolution, and obtaining a feature set; s2, extracting a mapping attention gate feature of the feature set and fusing the mapping attention gate feature with a gate control signal to obtain an attention gate control feature; s3, extracting multi-scale spatial features of the attention gating features, and aggregating and outputting the multi-scale spatial features; and S4, processing an output result in the step S3 through output convolution to obtain a segmentation result. The segmentation efficiency and accuracy are remarkably improved with fewer parameters and lower calculation requirements.

The invention provides an audio frequency measurement method and device, a medium and equipment, and the method comprises the steps: obtaining and sampling audio data, and obtaining a sampling sequence of N samples; in the sampling sequence, selecting a first sample set and a second sample set with different initial positions and the same number of samples; respectively performing fast Fourier transform (FFT) on the first sample set and the second sample set to obtain a first frequency spectrum sequence representation and a second frequency spectrum sequence representation, and respectively calculating phase angles of the first frequency spectrum sequence representation and the second frequency spectrum sequence representation at the maximum frequency spectrum amplitude to obtain a first phase angle and a second phase angle; and comparing the first phase angle with the second phase angle, selecting different FFT phase difference frequency estimation algorithms according to a phase angle comparison result, and calculating to obtain the audio frequency. The frequency measurement precision can be improved.

The invention relates to a rolling bearing fault diagnosis method. The invention relates to the technical field of rolling bearing fault diagnosis, and the method comprises the steps: decomposing a bearing signal through peak envelope spectrum Fourier decomposition to obtain a component signal, carrying out the envelope spectrum analysis of the component signal, and converting a component envelope spectrum into a gray-scale map as an input feature of a convolutional neural network for fault diagnosis. According to the method, the original bearing signal is decomposed by using peak envelope spectrum Fourier decomposition, a frequency band with rich fault information is effectively separated from a noise frequency band, the component signal is converted from a time domain into an envelope spectrum and further converted into a feature enhanced grey-scale map, the obtained envelope spectrum grey-scale map has obvious fault features, and the fault detection accuracy is improved. And the problem that a traditional fault diagnosis method needs complex feature extraction is avoided. Diagnosis experiments are carried out on bearing vibration signals of different fault types, and the result shows that the method is higher in diagnosis precision and higher in anti-noise capacity.

The embodiment of the present application provides an image encryption method, device, electronic device and readable storage medium, which belongs to the field of image processing technology. The image encryption method includes: determining the target area where the content to be encrypted is located in the original image to be encrypted; performing mask processing on the target area to determine the mask information of the target area; adding adversarial noise and auxiliary noise to the original image to generate a noise image; performing fusion processing on the original image, the noise image and the mask information to generate a target encrypted image.

The invention discloses a federated learning method based on a model deviation and variance distribution model, and relates to the field of machine learning. According to the method, the variance of the model uploaded by the client in each round is calculated on the server, the offset parameter of the model of each client relative to the global model is learned, the personalized information of each client is extracted, and a personalized initial model is provided for local learning of the client in each round. Through personalized model learning, the learned model has better performance on local data, the problem that existing personalized federated learning depends on extra overhead on the client is solved, extra data does not need to be transmitted between the client and the server, and privacy of client data is better protected. In addition, by means of hierarchical calculation of the model, calculation and storage expenses in the learning process are reduced, meanwhile, the performance of the personalized model is further improved, and the prediction accuracy of the data is improved.

The present invention discloses a continuous learning digital predistortion system and method for intelligent scenes, including a configuration adjustment module, a continuous learning digital predistortion module, a linearity check module, a training data storage module, a model coefficient extraction module and an objective function module; the configuration adjustment module changes the input signal configuration in real time, the input signal obtains a predistortion signal through the continuous learning digital predistortion module, the predistortion signal is transmitted to the power amplifier to obtain an output signal, the feedback loop collects the output signal, and enters the linearity check module to check the linearity. The present invention can greatly reduce the model coefficient update requirements of the digital predistortion technology in the intelligent scene, thereby reducing the application cost of the digital predistortion technology.

The invention discloses a semantic image transmission method and system based on a lightweight network. The method comprises the following steps: acquiring an input image; based on a Swin Transform architecture, quantization operation and non-linear operation integer approximation are introduced, and an image transmission network is obtained; the image transmission network comprises an encoder, a decoder and a channel modulation module; and transmitting the input image based on the image transmission network. The system comprises a memory and a processor used for executing the semantic image transmission method based on the lightweight network. By using the method, the calculation complexity and the energy consumption of the model are remarkably reduced, so that the method is more suitable for being deployed on edge equipment with limited calculation resources and energy consumption, and is more suitable for a high-speed image transmission scene. The method can be widely applied to the field of data transmission.

The invention discloses a water jet cutter path automatic planning method and system suitable for vegetable processing, and belongs to the technical field of food processing. The method comprises the following steps: firstly, carrying out image acquisition and analysis on vegetable slices, and identifying a vegetable slice area and a wood area of each vegetable slice from an image; in each vegetable slice, performing expansion treatment on the vegetable slice area and the wooden area, and subtracting the wooden area from the vegetable slice area to obtain a meat quality area; then discrete nodes are extracted from the contour line of the meat quality area, the number of the discrete nodes is reduced through fitting, and the fitted discrete nodes are recorded in a clockwise or anticlockwise sequence; and finally, generating a water jet cutter path of each vegetable slice according to the record sequence of the discrete nodes. The water jet cutter path automatic planning scheme is based on image recognition, planning time consumption is short, computing resource consumption is extremely low, and the efficient requirement of the vegetable processing industry can be met.

The present disclosure relates to the field of artificial intelligence technology, and in particular to a method, device and storage medium for a human-machine collaborative topic classification search mode, the method comprising: extracting a feature vector from an input query text, the feature vector comprising a plurality of input features; adjusting a preset sparsification model according to the feature vector to obtain an adjusted sparsification model, the adjustment comprising pruning a weight matrix of the sparsification model, the weight matrix comprising a plurality of weights, the weights being used to indicate the contribution of the corresponding input features to the output result of the sparsification model; according to the feature vector, obtaining the corresponding topic classification search result through the output of the adjusted sparsification model. The embodiment of the present disclosure dynamically adjusts and optimizes the sparsification model according to the feature vector of the query text, which involves dynamically pruning the weight matrix, thereby outputting the search results using the adjusted sparsification model to ensure the relevance and accuracy of the search results.

Embodiments herein provide a monitoring module (100) and a method for generating a signal indicative of a performance degradation of a capacitor (105) in an HVDC power converter (107). The method comprises sampling an input load signal of the capacitor using a measurement circuit (101). The method further comprises calculating a capacitance value of the capacitor using small signal analysis in a frequency domain using a frequency domain unit (102) and the measured input load signal. The method further comprises calculating a power loss factor of the capacitor using small signal analysis in a time domain using a time domain unit (103) on the measured input load signal. The method further comprises generating a signal indicative of the performance degradation of the capacitor based on the calculated capacitance value and the calculated power loss factor using a warning unit (104).

Implementations relate to encrypting audio data with utterance features generated from the audio data. Some of those implementations include generating an utterance feature from a portion of the audio data, encrypting at least a portion of the audio data using the utterance feature, and providing the encrypted audio data to one or more applications for decryption. The audio data is decrypted for further processing using speaker features previously generated from the utterance. Other implementations relate to receiving a speaker feature and comparing the speaker feature to an utterance feature generated from audio data. The audio data is provided to a target application that provides speaker features that match utterance features generated from the audio data.

The present invention relates to an environmental map construction method and a robot, which is characterized in that the environmental map construction is completed by the following steps: step 1, start the mapping and positioning module and the path planning module, and establish the automatic exploration module; step 2, generate a frontier position point queue D; step 3, determine whether the queue D is empty, if yes, end; if no, enter step 4; step 4, take out the first position point T in the queue D, determine whether the point T is reachable, if no, add the point T to the blacklist and return to step 3, if yes, enter step 5; step 5, move to point T, start the exploration timing and enter step 6; step 6, wait for n seconds, update the queue D, and move to point T; step 7, determine whether the current round of exploration timing has timed out, if yes, add T to the blacklist and return to step 2, if no, enter step 8; step 8, determine whether the robot has reached point T, if yes, return to step 2; if no, return to step 6. Compared with the prior art, the present invention reduces the calculation requirements during the exploration period.

The present application provides an ultrasonic color imaging method, system, apparatus, computing device and computer-readable medium. The method includes: receiving multiple ultrasonic receiving signals provided by an ultrasonic transducer, wherein the center frequencies of the multiple ultrasonic receiving signals are different from each other; dividing the multiple ultrasonic receiving signals into a preset number of groups, wherein the preset number is equal to the number of channels of a preset color mode; and using each group of ultrasonic receiving signals as a component of one of the channels of the preset color mode and generating a color image according to the preset color mode. The method generates a color image by using multiple ultrasonic receiving signals with different center frequencies as channel components of the color mode, which can make full use of the color perceptible space of the human eye to more completely and accurately reflect the information of the target object, and the required calculation amount is low, the hardware performance requirements are low, and it is more suitable for clinical applications.