33 results about "Trend detection" patented technology

The invention relates to the technical field of tunnel disaster identification, in particular to a tunnel portal icing disaster identification and prediction method based on a multi-modal data fusion technology. According to the technical scheme, the method comprises the steps of multi-source data acquisition, heterogeneous data processing, feature level fusion, disaster recognition, space-time prediction and dynamic early warning. According to the invention, a multi-modal sensor is deployed to collect tunnel portal temperature, space structure and environmental parameters, intelligent processing and multi-stage feature fusion are carried out, ice layer distribution identification, icing trend detection and time-space prediction are realized by using a deep network, an ice melting device is dynamically activated in combination with a graded early warning mechanism, and vehicle early warning is linked. A sensing, analysis, prediction, disposal and calibration closed loop is formed, the accuracy of ice coagulation disaster detection, the prospective performance of prediction and the intelligence of disposal are remarkably improved, and the tunnel traffic safety and the long-term robustness of the system are guaranteed.

The invention relates to the technical field of numerical control system fault pre-diagnosis, and particularly discloses a fault monitoring method and system for a numerical control machine tool. The method comprises the steps that a spindle vibration signal, a cutting force signal and a spindle current signal are collected; outputting an alignment signal set; respectively executing adaptive filtering processing according to the alignment signal set; constructing a multi-dimensional feature vector; constructing a health index under the processing behavior period; then accumulated trend deviation analysis is executed through a CUSUM algorithm; and judging whether the tool wear trend is abnormal. The technical problem that in the prior art, static threshold judgment is carried out based on a single signal source, and especially in a numerical control machining scene with signal noise interference, accurate early warning of the early wear trend of a tool cannot be achieved is solved. Due to the fact that a period calibration mechanism fusing spindle vibration, cutting force and current signals, feature vector entropy weight optimization and a CUSUM trend detection mechanism are adopted, the pre-diagnosis level of the numerical control machine tool on cutter state changes is improved.

The invention discloses a battery SOH dynamic prediction method based on federated learning and working condition adaptive fusion. The method comprises the steps that an end-side heterogeneous multi-modal feature extraction mechanism is adopted to extract physical features and working condition features of each battery from a battery changing cabinet end and a vehicle-mounted end; a double-branch fusion prediction network is constructed, and the double-branch fusion prediction network is trained through cooperation of a battery changing cabinet end global aggregation mechanism and a vehicle-mounted end self-adaptive mechanism; processing the extracted features by adopting two branch networks in the trained double-branch fusion prediction network and a gating fusion mechanism to obtain a predicted state of health (SOH) of each battery; the method comprises the following steps: constructing a predicted SOH sequence of each battery, constructing a multi-time-scale early warning module through a short-term mutation detection method and a long-term trend detection method, carrying out judgment operation on the predicted SOH sequence of each battery, judging whether each battery is in a risk state or not, and outputting an alarm when the battery is in the risk state; and the SOH prediction precision and the early warning capability of attenuation abnormity are improved.

The invention provides an event detection method and system for a mechanical overspeed test of a steam turbine generator unit, relates to the technical field of state monitoring and fault diagnosis of electric power equipment, and aims to solve the problems that existing mechanical overspeed test recognition is prone to false alarm and missing alarm, and positioning at a critical moment is not accurate. The method comprises the following steps: acquiring the active power of a generator and the rotating speed of a steam turbine, carrying out preprocessing and rolling smoothing processing to obtain a smoothing power and smoothing rotating speed sequence, determining a first key time point based on the smoothing power sequence, carrying out segmented backtracking and descending trend detection on the smoothing power sequence, and determining a second key time point. And carrying out descending track detection on the smooth rotating speed sequence, determining a third key time point, backtracking the smooth rotating speed sequence, determining a fourth key time point, determining a fifth key time point based on the original rotating speed, and judging whether a mechanical overspeed test is carried out or not. According to the invention, problems in the prior art are solved, and reliable and automatic identification and accurate event definition of the mechanical overspeed test are realized.

The invention discloses a time series data anomaly detection method and device, electronic equipment and a computer program product. The method comprises the following steps: acquiring a to-be-analyzed time sequence queue comprising a plurality of to-be-analyzed time sequence data; sliding a preset sliding window on the to-be-analyzed time sequence queue to obtain window data corresponding to the preset sliding window after each sliding; detecting a short-time trend of the window data according to a plurality of pieces of window data in the first window sequence; under the condition that the short-time trend shows that trend change occurs in the multiple pieces of window data in the first window sequence, the long-time trend of the window data is detected according to the multiple pieces of window data in a second window sequence, and the length of the second window sequence is larger than that of the second window sequence; and generating abnormal early warning information under the condition that the long-time trend shows that trend change occurs in the multiple pieces of window data in the second window sequence. The technical problem that in the prior art, the accuracy rate of an anomaly detection result obtained through trend detection is low is solved.

The present application relates to the technical field of lesion recognition, in particular to an automatic lesion recognition ultrasonic system for real-time monitoring, which comprises a delay path discrimination module, a gray level trend detection module, a texture feature screening module, a feature fusion sorting module and a region highlight labeling module. Based on continuous ultrasonic frames, the collected deep tissue echo path is analyzed, and the echo arrival time of each pixel point in the continuous frame is detected. The present application supports multi-type data fusion judgment through a comprehensive judgment process supported by multi-dimensional parameter collaborative screening, penetration behavior, gray level trend and texture aggregation. The regional abnormal priority sorting mode improves the hierarchy of lesion feature discrimination, provides partition directional recognition for local structural abnormalities and early micro-variation, and converts the feature judgment result into a high confidence region label by a weight aggregation method. The image output process automatically completes the real-time visual presentation of the high-risk area, improving the clarity and pertinence of the lesion region presentation.

The application discloses a cutterhead mud cake detection method and device, electronic equipment and a storage medium, and belongs to the technical field of tunnel construction. The detection method comprises the following steps: reading historical tunneling data from a PLC controller and an industrial computer of a shield tunneling machine; training a parameter trend detection model by using mud cake related parameters in the historical tunneling data; the parameter trend detection model comprises a one-dimensional convolution layer, an LSTM layer and a full connection layer; obtaining mud cake related parameters of the shield tunneling machine in a target time period, inputting the mud cake related parameters in the target time period into the parameter trend detection model, and obtaining a change trend of the mud cake related parameters in the target time period; judging whether the change trend of the mud cake related parameters in the target time period conforms to cutterhead mud cake characteristics; if yes, it is determined that the shield tunneling machine is in a cutterhead mud cake state; and if no, it is determined that the shield tunneling machine is not in the cutterhead mud cake state. The application can improve the accuracy of detecting cutterhead mud cake.

The abnormality analysis text automatic generation method, system and equipment of DHI interpretation report belong to the cross field of breeding technology and data text generation technology. In order to solve the problems of large workload and low efficiency of manually writing abnormality analysis text in the current DHI interpretation report, the DHI key performance index data is first acquired, the position of the data value of this month in the corresponding abnormality degree type array is located, the performance index name and the corresponding abnormality degree value are directly spliced, and the description text of the performance index static abnormality is obtained. Meanwhile, the global movement, local movement and standardization processing are carried out on the data of this month and the historical data, a single-layer bidirectional GRU network is used as an encoder for processing, the hidden states of all time steps and the hidden state of the decoder t-1 step are processed by using an attention layer, and an LSTM long short-term memory network is used as a decoder to generate the description text of the performance index dynamic trend detection.

The invention relates to the technical field of speech recognition and audio signal processing, in particular to an ultra-low power consumption speech wake-up system with adaptive threshold adjustment capability, which comprises an input sound processing module and an energy trend detection module for monitoring and receiving an audio signal x (t) from a microphone and outputting an envelope signal En (t); the adaptive control module updates the gain Gain of the programmable amplifier according to the envelope signal En (t); the programmable amplifier adopts the updated gain Gain to amplify the audio signal x (t) to obtain a signal x '' (t), and the signal x '' (t) passes through a band-pass filter to obtain a signal V 'BPF (t); the threshold judgment and wake-up output module adaptively adjusts a wake-up threshold Threshold according to the envelope signal En (t), and outputs a wake-up mark digital signal; the problem that an existing fixed sound pressure threshold value scheme cannot automatically adjust the wake-up threshold according to environmental noise changes is solved.

The invention discloses a production line visual management platform based on dynamic task data driving, and relates to the technical field of intelligent manufacturing and industrial informatization, and the platform comprises a dynamic task data collection module which is used for collecting the operation state data of production line stations, equipment, personnel and materials in real time; the task evolution modeling module is used for constructing a task evolution process according to the collected data; the task granularity self-adaptive refining module is used for dynamically adjusting the task management granularity according to the task state change; the abnormal fluctuation trend detection module is used for detecting abnormal fluctuation in the task state data; the visual management module is used for generating a production line visual interface based on the task evolution and the anomaly detection result; and the scheduling decision linkage module is used for optimizing a production line scheduling strategy according to the anomaly detection result and the task state. The problems of static data updating, task granularity fixing and anomaly response lagging existing in an existing production line visual management platform are solved.

The invention discloses a real-time anomaly detection and intelligent early warning system and method based on artificial intelligence, and relates to the technical field of artificial intelligence and data security, and the method comprises the steps: obtaining historical data assets, and constructing a tensor potential energy function; guiding the generated detector particles to migrate through the tensor gradient of the tensor potential energy function; negative selection is carried out on the migrated detector particles, an initial detector is generated, cloning is carried out on the initial detector, a clone detector is generated, and the detector is migrated to a potential abnormal region based on the tensor gradient; when the clone detector stably resides in the potential abnormal region, recording an evolution trajectory of the clone detector, and constructing a trend memory unit; according to a trend memory unit, generating a trend detector and migrating to a detector particle insufficient coverage area based on a tensor gradient; and performing response matching on the to-be-detected data assets, the clone detector and the trend detector to complete anomaly detection early warning. And high detection precision and early warning response rate are kept in a complex data environment.

Methods and systems for media trend detection and maintenance are provided herein. A set of media items each having common media characteristics is identified. A set of pose values is determined for each respective media item of the set of media items. Each pose value is associated with a particular predefined pose for objects depicted by the set of media items. A set of distance scores is calculated. Each distance score represents a distance between the respective set of pose values determined for a media item and a respective set of pose values determined for an additional media item. A coherence score is determined for the set of media items based on the calculated set of distance scores. Responsive to a determination that the coherence score satisfies one or more coherence criteria, a determination is made that the set of media items corresponds to a media trend of a platform.

The invention provides a measurement data change trend detection method and system and a computer storage medium. The measurement data change trend detection method comprises the steps that historical measurement data and new measurement data are integrated; sequentially selecting a plurality of groups of preset quantity of data from the integrated historical measurement data and the new measurement data according to the time sequence, and judging whether a second extreme value in each group of preset quantity of data is behind the first extreme value or not; if it is judged that the second extreme value in the preset number of data of the current group is behind the first extreme value, intercepting interval data from the first extreme value to the second extreme value in the preset number of data of the current group; and combining all the interval data according to the time sequence, and judging whether the change trend of the combined interval data is abnormal or not. According to the technical scheme, the change trend of the measurement data can be effectively detected.

The invention provides an etching end point detection method, a semiconductor structure and a manufacturing method thereof, and the etching end point detection method can accurately determine an etching end point based on the direct detection of the concentration of a first reaction product in an etching process. Compared with trend detection of the concentration of the first reaction product through a spectrum mode and the like, the accuracy of the etching end point determined in the method is higher. Meanwhile, according to the manufacturing method of the semiconductor structure based on the etching end point detection method, the etching end point can be automatically detected in the etching process of the wafer to be etched, etching is stopped, excessive etching or insufficient etching is avoided, the accuracy of the size of the formed semiconductor structure and the stability of the etching process are guaranteed, and the yield is improved. Therefore, the manufacturing yield of the semiconductor structure is improved.

This invention discloses a method for analyzing soil pollution change trends, comprising: acquiring a dataset of soil monitoring points at multiple time points within a target area; delineating core areas and time periods where pollutants in the soil exhibit significant upward or downward trends and spatially form a pattern; spatiotemporally matching the core areas and time periods with human activity factors and natural factors, constructing spatial statistical models, and obtaining the contribution rates of human activity factors and natural factors to the pollution change trend; this invention effectively integrates the inconsistency problem of spatiotemporal dynamic monitoring data, overcomes the data comparability barrier caused by changes in monitoring points, and realizes a closed-loop analysis from trend detection to cause analysis, significantly improving the accuracy and practicality of soil pollution source tracing and risk management.

The invention relates to the technical field of lesion recognition, in particular to an automatic lesion recognition ultrasonic system for real-time monitoring, which comprises a delay path judgment module, a gray scale trend detection module, a texture feature screening module, a feature fusion sorting module and a region highlight labeling module. The acquired deep tissue echo path is analyzed, and the echo arrival time of each pixel point in continuous frames is detected; according to the method, multi-type data fusion judgment is supported through a comprehensive judgment process of multi-dimensional parameter collaborative screening, penetration behavior, gray trend and texture aggregation, hierarchy of lesion feature judgment is improved through a regional anomaly priority sorting mode, and partition directional recognition is provided for local structure anomaly and early tiny variation; the weight aggregation method converts a feature judgment result into high-confidence region labeling, real-time visual presentation of a high-risk region is automatically completed in the image output process, and the presentation definition and pertinence of a lesion region are improved.

This invention discloses a clean air leak detection system and method based on the Internet of Things (IoT), relating to the field of industrial process monitoring. The method includes: preprocessing the acquired raw monitoring signal to obtain a standardized signal; performing robust statistical analysis on the standardized signal based on a sliding window to obtain the time-varying baseline center and the time-varying baseline fluctuation intensity; correcting the standardized signal according to the time-varying baseline center to obtain a residual signal; performing wavelet multi-resolution decomposition on the residual signal to obtain approximation coefficients and detail coefficients distributed at different scales; performing threshold shrinkage processing on the detail coefficients and reconstructing a trend enhancement signal using the processed detail coefficients and approximation coefficients; adaptively setting detection parameters based on the time-varying baseline fluctuation intensity, and using a sequential cumulative sum algorithm to detect persistent shift trends in the trend signal. Through adaptive baseline estimation, wavelet denoising, and sequential cumulative sum detection, high sensitivity and low false alarm detection of weak leak trends are achieved.

The invention discloses an access method and system on an industrial data link, and relates to the field of industrial data processing, and the method comprises the steps: collecting the time sequence monitoring data of industrial equipment, deploying a trend detector at the edge side, and dynamically adjusting the anchoring frequency on the link according to the data sudden change intensity and the abnormal confidence; when a transient event is captured, an event snapshot containing a time window and an associated point location is generated, the event snapshot is linked preferentially through an independent channel, and the event snapshot falls back to a conventional anchoring rhythm after abnormities subside. When historical data is archived in a roll-by-roll mode, connection records containing inter-roll associated indexes are generated in roll switching to be linked, a cross-roll index directory corresponding to a hash root on the chain is maintained, and an anchoring time period window on the chain is inquired before roll cutting of an archiving system so as to align a boundary. According to the method, the problems of transient abnormal tracing blind areas and cross-volume chain breakage caused by existing fixed anchoring are solved, the data tracing continuity and the key event capturing capability are improved, and reliable support is provided for industrial data auditing and fault tracing.

The application belongs to the technical field of pipeline leakage detection, and particularly relates to a pipeline pressure leakage signal enhancement method, which comprises the following steps: obtaining upstream and downstream pipeline pressure signals; performing trend detection and trend elimination on the upstream and downstream pipeline pressure signals to obtain the upstream and downstream pipeline pressure signals after trend elimination; performing multi-scale characterization on the prominence degree of interval sub-signal amplitude in the upstream and downstream pipeline pressure signals after trend elimination to obtain the prominence degree characterization function of each interval sub-signal in the upstream and downstream signals and the optimal filtering parameter; and multiplying the upstream and downstream pipeline pressure signals filtered by the optimal filtering parameter with the prominence degree characterization function of each interval sub-signal in the upstream and downstream signals to obtain the enhanced upstream and downstream pipeline pressure signals. The method solves the problem of weak leakage signal enhancement and ensures the optimization of the leakage signal enhancement effect.

The invention provides a remote sensing image target trend detection method which can be applied to the technical field of remote sensing image processing and analysis. The method comprises the steps that an input image is acquired and preprocessed; based on spectral features and geographic position prior information, a semantic segmentation network is adopted to separate a target area from the preprocessed image; constructing a cost measurement function by using an optimal transmission theory, solving an optimal transmission plan, and obtaining feature mapping after image smoothing; based on feature mapping, texture features and motion features are extracted, and spatial-temporal features of the target area are obtained through fusion; capturing a space-time dependency relationship in the multi-temporal remote sensing image feature sequence by adopting a time sequence analysis model; constructing a target trend prediction model based on the feature mapping, the spatio-temporal features and the spatio-temporal dependency relationship; and outputting a trend prediction result by using the target trend prediction model. The invention further provides a remote sensing image target trend detection device, equipment and a medium.

The present application relates to the technical field of tunnel disaster identification, and particularly relates to a tunnel portal ice disaster identification and prediction method based on multi-modal data fusion technology. The technical scheme comprises multi-source data acquisition, heterogeneous data processing, feature-level fusion, disaster identification, spatio-temporal prediction and dynamic early warning. The present application collects tunnel portal temperature, spatial structure and environmental parameters by deploying multi-modal sensors, and realizes ice layer distribution identification, icing trend detection and spatio-temporal prediction by intelligent processing and multi-level feature fusion using a deep network. The present application dynamically activates the ice-melting device and links the vehicle warning by combining the hierarchical early warning mechanism, forms a closed loop of perception, analysis, prediction, disposal and calibration, significantly improves the accuracy of ice disaster detection, the foresight of prediction and the intelligence of disposal, and guarantees the safety of tunnel traffic and the long-term robustness of the system.

This invention discloses a method, apparatus, electronic device, and storage medium for processing multipath effects in a Time-of-Flight (TOF) sensor. After acquiring the elevator door closing distance and the initial TOF image collected by the TOF sensor, the method identifies the target image region affected by the multipath effect from the initial TOF image based on the door closing distance. The target image region is then masked from the initial TOF image to obtain a target TOF image. Trend detection is performed based on the target TOF image. This allows for the identification of the target image region affected by the multipath effect caused by the door frame edge or door panel edge from the TOF image at different stages of elevator door closing, based on the door closing distance. Furthermore, the target image region is masked from the TOF image, thus excluding the image region with multipath effect from trend detection. This avoids the influence of multipath effect on trend detection, improves the accuracy of trend detection, and makes controlling the elevator door closing or opening more safe and reliable based on the trend detection results.

The invention provides a model training method, a device trend detection method, a device and equipment. Relates to the technical field of safety management. The method comprises the following steps: collecting multiple pieces of historical data of power equipment to obtain a first power data set; modifying and / or deleting data in the first power data set to obtain a second power data set; taking the second power data set as input of an initial learning module, taking output of the teacher model as a training target of the initial learning module, and adjusting corresponding parameters in the initial learning module to obtain a feature learning module; extracting a first fusion feature corresponding to each piece of historical data in the second power data set through a feature learning module; based on the first fusion feature, training an equipment trend detection module to obtain a trained equipment trend detection module; and integrating the feature learning module and the equipment trend detection module to obtain an equipment trend detection model. Through the method and the device, the effect of improving the equipment trend identification accuracy is achieved.

The invention belongs to the technical field of intelligent fault detection of an airport production operation system, and discloses an airport production operation system fault identification method and device based on machine learning. The method comprises the following steps: preprocessing collected service request information of a production operation system recorded in a system log; sudden change detection and trend detection are carried out on the airport production operation system faults, and sudden change abnormity of the airport production operation system faults and trend abnormity of the airport production operation system faults are detected respectively; the method is applied to airport production operation scenes, and the recognition rate, the false alarm rate and the response time index are counted respectively. According to the method, the identification rate of the abnormal events of the index mutation class is 98.7%, and the false alarm rate is lt; and the recognition rate is 91.2% within 7 days for continuous abnormity of the trend. The application effect is obvious, the abnormal state of the system can be accurately detected in advance, and timely processing is carried out to avoid expansion of the fault range.

The application discloses a real-time anomaly detection and intelligent early warning system and method based on artificial intelligence, relates to the technical field of artificial intelligence and data security, and comprises the following steps: acquiring historical data assets and constructing a tensor potential function; guiding the migration of generated detector particles through the tensor gradient of the tensor potential function; filtering the migrated detector particles through negative selection to generate an initial detector, performing cloning on the initial detector to generate a cloned detector, and migrating to a potential abnormal area based on the tensor gradient; when the cloned detector stably resides in the potential abnormal area, recording the evolution track of the cloned detector and constructing a trend memory unit; generating a trend detector according to the trend memory unit and migrating to an area where the detector particles are insufficient based on the tensor gradient; and performing response matching on the to-be-detected data assets, the cloned detector and the trend detector to complete anomaly detection and early warning. The application ensures high detection accuracy and early warning response rate in a complex data environment.

The application provides a cotton picker picking head clogging trend detection and early warning method and system based on double-path feature fusion and prediction reconstruction, comprising obtaining picking head vibration signals and cutting into vibration windows; statistical physical features and original vibration sequence deep features of each window are extracted and window-level fusion is performed to construct a time series context feature sequence; the time series context feature sequence is input into a time series prediction model to obtain prediction target features and hidden states of a future target window; hidden space distribution parameters are generated based on the hidden states and latent variables are determined, and reconstruction target features are obtained through a reconstruction constraint model; a window anomaly score is constructed according to a prediction error, a reconstruction error and a hidden space divergence, and an initial healthy window is combined to establish an intra-source adaptive baseline; when the anomaly score exceeds a warning threshold and the continuous window condition is met, a warning result is output. The application considers feature interpretability and deep representation ability, realizes high-precision early warning of clogging trend anomalies, and is self-adaptive, strong and highly stable.

In general, devices, systems, and techniques are described for determining periodicity metrics for electrical signals sensed from a patient. In one example, a system includes processing circuitry configured to receive, from sensing circuitry, a plurality of electric signals from a patient sensed over a time duration, determine characteristic values for the plurality of bioelectric signals, and determine, based on the characteristic values, a periodicity metric indicative of one or more repeating periods for the plurality of bioelectric signals. The processing circuitry can then control delivery of electrical stimulation therapy based on the periodicity metric.

A user information generation method and apparatus is proposed. The method may include collecting, from at least one smart device counting apparatus, smart device count information generated using a signal transmitted by at least one smart device within a predetermined range. The method may also include generating user count information as a result of calculating a number of users within the range by applying an average number of smart devices possessed by each person to the smart device count information. The method may further include generating location estimation information about the smart device, based on the smart device count information, and generating movement trend detection information about users based on the location estimation information about the smart device, the user count information, and map information.

The application discloses an industrial data chain access method and system, and relates to the field of industrial data processing.The method comprises the following steps: collecting time sequence monitoring data of industrial equipment, deploying a trend detector on the edge side, dynamically adjusting the anchor frequency on the chain according to the data mutation intensity and abnormal confidence, generating an event snapshot containing a time window and associated point positions when capturing a transient event, preferentially uploading the event snapshot to the chain through an independent channel, and falling back to the normal anchor rhythm after the abnormality subsides.When archiving historical data, a link record containing an inter-volume association index is uploaded to the chain during volume switching, a cross-volume index directory corresponding to the hash root on the chain is maintained, and the anchor time period window on the chain is queried before the archiving system switches the volume to align the boundary.The application solves the problems of transient abnormality tracing blind area and cross-volume disconnection caused by the existing fixed anchor, improves the data tracing continuity and key event capturing capability, and provides reliable support for industrial data auditing and fault tracing.

This invention provides a method for detecting target trends in remote sensing images, applicable to the field of remote sensing image processing and analysis. The method includes: acquiring and preprocessing an input image; separating the target region from the preprocessed image using a semantic segmentation network based on spectral features and prior geographic location information; constructing a cost metric function using optimal transport theory, solving for the optimal transport plan, and obtaining a smoothed feature map of the image; extracting texture and motion features based on the feature map, and fusing them to obtain the spatiotemporal features of the target region; capturing the spatiotemporal dependencies in the feature sequences of multi-temporal remote sensing images using a time series analysis model; constructing a target trend prediction model based on the feature map, spatiotemporal features, and spatiotemporal dependencies; and outputting the trend prediction result using the target trend prediction model. This invention also provides a remote sensing image target trend detection device, equipment, and medium.