82results about How to "Real-time processing" patented technology

The invention discloses a speech recognition streaming feature extraction and batch reasoning method for a high-concurrency artificial intelligence real-time dialogue system. According to the method, a heterogeneous computing architecture in which a CPU (Central Processing Unit) and a GPU (Graphics Processing Unit) cooperatively work is adopted, operations such as preprocessing, feature extraction and normalization of voice data are executed through the CPU, and batch feature reasoning is executed by utilizing the GPU; the method specifically comprises the following steps: after segmenting and preprocessing voice data at a CPU (Central Processing Unit) end, executing Mel filter bank feature extraction and cepstrum mean variance normalization to obtain a standardized acoustic feature vector; extracting feature vectors from the feature buffer area in batches at the GPU end, filling and aligning the feature vectors, and sending the feature vectors to a neural network model for parallel reasoning calculation; according to the method, streaming processing and batch reasoning in the speech recognition process are realized, and the concurrent processing capability and the real-time performance of the system are effectively improved.

The invention discloses a water and soil conservation monitoring system and method based on the Internet of Things technology, and relates to the technical field of water and soil conservation monitoring. A data acquisition module of the Internet of Things technology is utilized to judge small, medium and large drainage basins according to the drainage basin length of a monitoring area; then real-time data of soil humidity, drainage basin correlation and water and soil conservation measures are collected in real time from a monitoring area, a calculation and evaluation module is used for respectively inputting, sequentially calculating and outputting a soil erosion rate TQ, a drainage basin monitoring area evaluation value LY and a water and soil conservation measure adjustment value SBF, and the water and soil conservation measure adjustment value SBF is calculated and output on the basis of the water and soil conservation measure adjustment value SBF; according to the method, the accuracy of monitoring and the pertinence to small, medium and large drainage basins are ensured, a closed-loop feedback mechanism is formed, an innovative monitoring method is realized, and the beneficial effects provide more scientific, comprehensive and effective support for water and soil conservation work.

The invention relates to the technical field of intelligent traffic, and particularly discloses a pedestrian movement track prediction method and system based on visual information and a storage medium, and the method comprises the following steps: unmanned aerial vehicle data collection, coordinate projection transformation, coordinate sampling and track aggregation, sampling of pedestrian ground real coordinates after projection transformation, and tracking of the pedestrian movement track. Aggregating the coordinate data into trajectory information of a plurality of pedestrians; spatial visual information and time information are combined, a global interactive directed topological graph is constructed, then spatial feature and time feature extraction is performed, time sequence features of pedestrian motion are efficiently extracted through a multilayer convolution structure, and finally the future motion trail of pedestrians is predicted. Through the method, pedestrian motion in a complex scene can be predicted more efficiently, and the application effect of the model in pedestrian flow management and public safety planning is enhanced. The future movement condition is predicted by analyzing the historical trajectory of pedestrians, the pedestrian flow management is obviously optimized, and the safety level of public places is improved.

The invention discloses a key area intelligent monitoring system based on communication towers, and relates to the field of communication technology, IoT and intelligent monitoring, and the key area intelligent monitoring system is characterized in that each communication tower is provided with a front-end sensing layer; the front-end sensing layer is used for acquiring sensor data of a key area and preprocessing the sensor data to generate preprocessed data; the channel transmission layer is used for transmitting the preprocessed data to the data processing and analyzing layer; the data processing and analysis layer is used for generating a self-adaptive abnormal behavior recognition model based on a machine learning model library, processing the preprocessed data according to the self-adaptive abnormal behavior recognition model, analyzing abnormal behaviors and generating a data analysis result; the application service layer is used for transmitting the data analysis result to the mobile terminal and formulating an early warning strategy according to abnormal behaviors in different scenes; the early warning strategy comprises an early warning threshold value and a response rule. According to the invention, comprehensive coverage of a monitoring area, rapid monitoring response and timely data processing can be realized.

The invention discloses a data security management and control method and device based on classification and grading, a medium and a product, and relates to the technical field of data security management, and the method comprises the steps: obtaining a data confidentiality grade labeling strategy; performing grade labeling on the data stored in each local computer and the data center according to a grade labeling strategy; establishing a data classification grading and security control model based on a neural network by adopting a transverse federated learning architecture; performing data set division based on the desensitized data, different from the data confidentiality level of the local computer, in the data center and the data in the local computer to obtain a training set and a test set; according to the training set and the test set, training the data classification and grading and security management and control model to obtain a trained model; and based on the trained data classification and grading and security management and control model, carrying out classification and grading on data with unknown confidentiality levels. According to the invention, accurate classification and security control can be carried out on the data with unknown confidentiality levels.

The present invention relates to the technical field of sleep diagnosis, and in particular to a sleep staging method based on deep learning, the method comprising: collecting a number of target physiological signals, preprocessing each target physiological signal to obtain a number of standard physiological signals; synchronously aligning and integrating each standard physiological signal in turn to obtain a standard sleep data set; determining the actual correlation between each physiological signal to be integrated in the standard sleep data set, and determining an actual signal combination strategy according to the actual correlation; generating an actual sleep data set according to the actual signal combination strategy, and extracting features from the actual sleep data set to generate actual feature parameters; inputting the actual feature parameters into a deep learning model to determine the actual sleep state; analyzing the actual sleep state, and performing corresponding sleep intervention according to the analysis results. The present invention realizes low-cost non-intrusive sleep monitoring, thereby improving people's physical health.

The invention discloses a high-speed denoising method and device suitable for an airborne area array Geiger avalanche diode radar, and the method comprises the steps: enabling a radar image collection control module to achieve the data collection of the area array Geiger avalanche diode radar, and carrying out the segmentation of effective pixels and noise pixels of a collected 64 * 64 image; self-adaptive denoising is added by analyzing a radar equation and equipment parameters; automatically setting a signal threshold according to avalanche diode data distribution in a single frame; the signal proportion is increased by using matched filtering; signals are extracted and noise is removed by using a time and space correlation method. According to the invention, real-time de-noising processing is carried out on acquired radar data by combining the inherent characteristics of the area array Geiger avalanche diode radar, data distribution in a single frame, a time-space correlation method and a matched filtering algorithm, so that the problem of data submerging of the Geiger avalanche diode radar under an extremely low synthesis frame number is solved; the background noise interference in the image can be effectively reduced, and the pixel signal centroid can be quickly obtained.

The invention discloses an image sensor detection method, an image dynamic detection method and a correction method, and the detection method comprises the steps: screening any abnormal pixel point in abnormal cluster points as a center, the number of the second bright spots in the rectangular array area containing the two layers of outer neighborhood single-pixel rectangular rings is smaller than the number threshold value, and the corresponding abnormal pixel points serve as dead pixels; wherein the abnormal pixel point is confirmed by judging whether the gray value of the corresponding pixel channel is greater than or equal to the abnormal pixel point reference threshold value; and screening the corresponding abnormal pixel points of which the outer neighborhood single-pixel annular region is less than three continuous second bright points by taking the corresponding abnormal pixel point reaching the preset number threshold as the center as defective pixels. According to the method, the spatial distribution characteristics of the positive samples and the negative samples are effectively utilized, the positive samples and the negative samples are precisely discriminated in combination with the spatial distribution statistical information, and a reliable correction basis is provided for dead pixel correction.

The invention discloses a target detection-based oiling machine hose production quality detection method. The method comprises the following steps of: 1) acquiring an image; 2) preprocessing the data; (3) the YOLOv10 model is trained and updated, an AncorBox and a four-scale detection network are redesigned, a Focal Loss adjustment loss function is introduced, and a new YOLOv10 model is obtained; then, the preprocessed data are substituted into the updated YOLOv10 model to be detected; and (4) outputting a detection result, and when detecting that a novel hose or a new defect type appears on the production line, training and updating the model in the step (3) again by the system by collecting new sample data. By means of the method, the technical problems that in the prior art, the detection efficiency is low, accuracy is not high, and universality is poor are solved, and the oiling machine hose production quality detection method can improve the detection accuracy, so that the product quality is improved, and the production cost is reduced.

The invention discloses a temperature control system for an inductor production line, which comprises a PLC (Programmable Logic Controller) part, a touch screen part, a temperature control part and a power switch part, and is characterized in that one side of the PLC part is connected with the touch screen part through an RS-485 communication interface; one side of the PLC part is connected with the temperature control part through an RS-485 communication interface, and the power switch part is connected with the PLC part, the touch screen part and the temperature control part. According to the temperature control system for the inductor production line, temperature control is accurate, the high-precision temperature sensor is adopted, and it is ensured that temperature data are accurately obtained in real time. A PID control algorithm and a temperature controller are used for processing temperature data in real time, operation is easy and convenient, a touch screen provides a visual man-machine interaction interface, heating and cooling equipment is automatically adjusted according to the preset temperature, manual intervention is reduced, production efficiency is improved, and energy consumption is reduced.

The invention belongs to the technical field of mining monitoring equipment, and particularly relates to a mining integrated multi-gas monitoring device and method and a storage medium, the mining integrated multi-gas monitoring device comprises a protective shell, and a multi-gas monitoring alarm instrument, a plurality of sucking pumps, a camera and a video display screen comprising a video processor are arranged in the protective shell. An electric rotating baffle is controlled to block the communication between the second gas conveying pipeline and the multi-gas monitoring alarm instrument, or a sucking pump is controlled to be started, and the electric rotating baffle is controlled to block the communication between air outside the goaf and the multi-gas monitoring alarm instrument; the monitoring device is directly installed on the mine sealing wall, the working face corner and / or the tunneling head-on position, continuous and automatic monitoring of the gas environment in the mine is achieved, a worker does not need to carry various instruments on site for operation, personal errors are reduced, and the data accuracy is improved.

The utility model discloses a coal mining machine roller wear monitoring and early warning device, which belongs to the technical field of coal mining machinery monitoring, and comprises a pair of monitoring semi-rings, a plurality of non-contact displacement sensors and a control module, the two side ends of each monitoring semi-ring are fixedly provided with mounting components, the pair of monitoring semi-rings is assembled into a monitoring ring through the mounting components, and the non-contact displacement sensors are arranged on the monitoring ring. A plurality of uniformly distributed monitoring holes are formed in the inner end of the monitoring semi-ring, the control module comprises a power supply and a microprocessor, the power supply is fixedly installed at the outer end of the monitoring semi-ring, the power supply and the non-contact displacement sensor are electrically connected with the microprocessor, and a plurality of sound-light alarm lamps are further installed at the outer end of the monitoring semi-ring. By adopting the non-contact displacement sensor and the laser measurement technology, the radial displacement change of the surface of the roller caused by abrasion can be monitored in real time, and data processing and abrasion degree calculation are performed through the microprocessor, so that the measurement accuracy and reliability are improved, and the measurement precision is improved. And errors caused by human factors in a traditional monitoring method can be avoided.

The invention particularly relates to an airborne high-resolution video transmission method. The method comprises the following steps: acquiring a to-be-processed image by a first device end, and preprocessing the to-be-processed image according to image format information of the to-be-processed image to acquire a first image; the first device end carries out image segmentation processing on the first image to obtain a plurality of image slices and sends the image slices; the second equipment end receives the image slices and performs image splicing processing to obtain a second image; carrying out image coding processing on the second image to obtain image coding data; the second equipment sends the image coding data to the target terminal; the first equipment end is a field programmable gate array (FPGA); and the second equipment end is a system chip SOC. According to the method, the standard format image and the non-standard format image can be processed in real time. The image slices are spliced at the SOC end, the slice image is restored to the original image, video compression coding of the whole image is carried out, and the problem that the edge of the compressed slice image is blurred and uneven is avoided.

The present invention provides a method for estimating the early fire area in a forest fire emergency rescue scenario, which relates to the field of fire technology. The method includes: in the early stage of forest fire emergency rescue, using a panoramic camera mounted on an unmanned aerial vehicle to collect fire images, and performing denoising and thin smoke removal on the panoramic images; obtaining pre-disaster satellite remote sensing images according to the location of the forest fire, and performing orthorectification and cropping processing; using the processed pre-disaster satellite remote sensing images as reference images, registering the processed post-disaster unmanned aerial vehicle panoramic images; comparing the pre-disaster remote sensing images with the registered post-disaster unmanned aerial vehicle panoramic images, extracting the fire area, and counting the fire area. The present invention solves the problems of field of view obstruction, estimation efficiency and estimation accuracy not being compatible in the existing fire area estimation methods, and meets the actual needs of forest fire emergency rescue.

The application discloses an Internet of Things mass service bill processing system, method and device and a storage medium, and comprises an application software service layer, a platform environment service layer and an infrastructure service layer.The application software service layer is used for collecting first service bills, and performing format analysis, deduplication, merging and distribution processing on the first service bills to obtain first bill files and first bill messages, and then uploading the first bill files and the first bill messages.The platform environment service layer is deployed with a distributed file system, a distributed coordination service, a data database and a streaming processing framework, is used for receiving the first bill files and the first bill messages, and performing full-amount synchronization on the data database according to the first bill files and incremental synchronization on the data database according to the first bill messages.The infrastructure service layer comprises a bottom cloud host, and is used for providing a computing resource pool, a storage resource pool and a network resource pool.The application improves the processing efficiency of service bills, optimizes the use perception of users, and can be widely applied to the technical field of data processing.

This invention provides a multi-channel image sensor data transmission system and optoelectronic system based on FPGA-PCIe, belonging to the technical field of video processing and acquisition systems. The multi-channel image sensor data transmission system includes an FPGA circuit board, external memory, and a back-end image processing platform. The FPGA circuit board includes an integrated image processing module, a bus control module, and a PCIe interface control module. Utilizing an FPGA to process and acquire multiple sensor video streams saves hardware costs and circuit board area. It also ports various operation operators to the FPGA, enabling real-time processing and acquisition of the currently transmitted video stream, improving the real-time performance of the operation. Simultaneously, it achieves control and transmission of multiple video streams, meeting the requirements for data transmission agility and high speed. This solves the problem in existing methods where inconsistent video resolutions and frame rates from different image sensors lead to insufficient data bandwidth to meet protocol requirements, resulting in the inability of PCIe devices at both ends of the FPGA to communicate normally.

The application relates to the field of high-speed communication and information network technology, in particular to a space expandable photoelectric information processing device, which comprises a device body, the device body comprises a power board module, an information processing board module, an interface board module, an expansion information processing module and a processing board heat dissipation top cover, a shielding block is arranged in the interface frame, and the device body comprises the processing board heat dissipation top cover. The application can not only improve the complex processing technology, overcome the deficiencies of the existing technology in the aspects of efficiency, cost, precision, material utilization rate, environmental protection and the like, improve the modularization and expandability of the system, improve the photoelectric data processing capacity, realize high integration and small and light design, help improve the heat dissipation performance and environmental adaptability of the device, and help unify the information processing of multiple types of photoelectric loads.

The present invention discloses a method for processing a target material distribution image in a coherent imaging system, and belongs to the technical field of coherent imaging image processing. The method includes configuring a reference laser interference signal subdivision multiple so that the exposure trigger frequency is more than 2 times higher than the highest frequency of the light interference signal in the observation band; designing a special narrow-band bandpass FIR filter according to the characteristic absorption wavelength of the target material component to be measured, the reference laser wavelength, and the reference laser interference signal subdivision multiple; based on the area array coherent imaging technology, acquiring the interference signal data cube image frame by frame according to the exposure trigger frequency; sending the newly added frame image data of the data cube to the FIR filter for parallel real-time processing to obtain a two-dimensional normalized image of the material distribution; and obtaining a two-dimensional concentration distribution image with quantitative concentration based on the background removal processing of the normalized image. The present invention realizes synchronous calculation in the process of acquiring the coherent imaging data cube frame image, saves storage resources, and improves the material distribution imaging rate.

The invention discloses a gas-liquid erosion experimental device for a U-shaped pipeline. The gas-liquid erosion experimental device comprises a gas phase injection system, a liquid phase injection system and an erosion device which are connected with the pipeline, the gas phase injection system is used for injecting gas into the pipeline; the liquid phase injection system is used for injecting an ammonium chloride solution into the pipeline, and an atomizing nozzle is assembled at the tail end of the liquid phase injection system; a U-shaped sample tube for carrying out an erosion experiment is assembled in the erosion device; a data collecting and processing system is assembled on the erosion device; the detection precision is high, and the injection conditions of a gas phase and a liquid phase can be accurately controlled through the gas phase injection system and the liquid phase injection system; the applicability is wide, and the influence on pipeline erosion under different ammonium chloride solution concentrations and air flows can be tested; operation steps are simple, phenomena are obvious, direct observation is facilitated, and experiment efficiency is improved.

A device and method for detecting the pull-out force of planetary gear shafts in automatic transmissions belong to the technical field of automotive automatic transmissions. The device includes a base, a gantry bracket, replaceable testing support fixtures, a pressure detector, a testing platform, a pressure sensor, a hydraulic cylinder, replaceable pressure heads, guide columns, and a hydraulic mechanism. All components are rigidly connected. The fixtures and pressure heads are interchangeable to accommodate different specifications of planetary gears. The hydraulic mechanism supplies pressure to the hydraulic cylinders, and the pressure sensor collects force values. The testing method includes four steps: fixture adjustment, placement and positioning, pull-out detection, result judgment, and completion. Hydraulic force is applied to press the shaft, the sensor collects data, and the detector performs judgment, storage, and data export. This invention, through multi-system collaborative design, achieves accurate and quantitative detection of the pull-out force at the riveting joint of the planetary gear shaft. It combines versatility, convenience, and data traceability, effectively solving many shortcomings of existing testing methods and providing reliable assurance for riveting quality control.

The invention relates to the technical field of punching machines, in particular to a CCD servo hydraulic sliding table duplex punching machine which comprises a punching machine body, a dust sticking mechanism is fixedly mounted on one side of the punching machine body, a fixing frame is fixedly connected to one side of the dust sticking mechanism, a mounting frame is arranged above the fixing frame, and a plurality of groove dust sticking rollers are arranged on the mounting frame in a sleeving mode. The mounting frame is arranged, the groove dust-sticking rollers corresponding to the strip-shaped grooves in the surface of the product in size and number are arranged on the mounting frame, impurities in the strip-shaped grooves in the surface of the product can be treated through the groove dust-sticking rollers, omission and dead angles are reduced due to the fact that the groove dust-sticking rollers can be accurately matched with the grooves, and therefore the cleaning efficiency is greatly improved; and meanwhile, the detachable mounting frame is arranged, and the movable abutting frame is arranged in the mounting frame, so that the groove dust-sticking rollers of different numbers and sizes can be conveniently limited, the groove dust-sticking rollers can be easily replaced or adjusted according to production requirements, the device is suitable for products of different shapes and sizes, and the flexibility of the device can be improved.

The invention relates to the technical field of medical image processing, in particular to a percussion frame recognition method, device and equipment. The method comprises the following steps: determining received N continuous image frames as input data in a process of receiving image frames sequentially transmitted by image acquisition equipment according to a time sequence; wherein N is greater than or equal to 2; inputting the input data into a feature extraction module to obtain N image features; inputting the N image features into a self-attention module to obtain N attention vectors; inputting the N attention vectors into a feature dimension reduction module to obtain a prediction probability that each image frame is a percussion frame; determining the image frame corresponding to the prediction probability meeting the preset standard in the N prediction probabilities as a percussion frame; the technical problem that an existing percussion frame determination method is low in determination efficiency can be solved.

This invention relates to a method for synchronous detection of road surface anomalies based on hyperspectral imaging, belonging to the field of hyperspectral image processing and remote sensing detection technology. It solves the problems of low data processing efficiency, limited detection functions, and insufficient real-time performance in existing technologies. The method includes: acquiring hyperspectral images of various road surfaces, extracting key features, generating hyperspectral fingerprints, labeling road surface categories, and constructing a hyperspectral fingerprint database; training a fingerprint comparison model based on the hyperspectral fingerprint database to obtain a pre-trained fingerprint comparison model; acquiring road hyperspectral images in real time, extracting key features, generating a hyperspectral fingerprint to be tested, and obtaining the road surface category of the hyperspectral fingerprint to be tested based on the hyperspectral fingerprint database, the pre-trained fingerprint comparison model, and a hierarchical comparison strategy. This enables simultaneous detection of multiple anomalies such as icing, water accumulation, and road surface deterioration on an UAV-borne terminal.

The invention discloses an Internet of Things data processing method and system based on data service, and a terminal. The method comprises the following steps: collecting initial equipment data of a specific type or a specific purpose of target Internet equipment in a preset time period; performing data cleaning and data induction on the initial equipment data to obtain target equipment data; performing real-time processing on the target equipment data to generate a data real-time analysis result and decision support; and carrying out deep analysis processing on the target equipment data to generate a deep analysis result, and providing a targeted optimization suggestion according to the deep analysis result. Real-time data cleaning and analysis functions are provided, the data quality and accuracy are improved, a large amount of data is rapidly processed and analyzed according to the real-time requirement under the Internet of Things scene, data in the Internet of Things environment can be rapidly processed in real time, and accurate analysis results and decision support are provided.

This utility model discloses a safety performance testing device for new energy vehicles, including an electric platform cabinet. A display screen is electrically connected to a built-in control board near the top of the front of the electric platform cabinet to intuitively display the results of different tests. A function key area is fixedly connected to the top of the display screen on the front of the electric platform cabinet, allowing operation of the entire safety performance testing program using multiple buttons within the function key area. A resistance tester is installed at the bottom of the inner part of the electric platform cabinet to test the resistance values ​​of the internal circuits. This utility model integrates multiple functions such as fast charging testing, slow charging testing, resistance testing, and potential equalization testing into a single electric platform cabinet, enabling the completion of core safety performance testing of new energy vehicles without the need for multi-station transfers. This significantly shortens the testing process, improves testing efficiency, and reduces equipment footprint and investment costs.

The present invention discloses a system for identifying complex pollutants in industrial agglomerations based on big data, and relates to the technical field of industrial pollutants. The system comprises an acquisition module, a processing module and an output module, wherein the acquisition module stores a basic information data set of an industrial park, a chemical knowledge information data set, a factory emission information data set and an emission index information data set. The efficient collaboration of the acquisition module, the processing module and the output module of the present invention realizes the comprehensive monitoring and accurate identification of complex pollutants in industrial agglomeration areas. The acquisition module integrates multi-source data and provides a rich basis for analysis. The predictive analysis model in the processing module uses a long short-term memory network to extract key features, and through continuous optimization and updating, ensures the accuracy and timeliness of the model, can accurately identify the conditions for the generation of complex pollutants, predict their development trends, and provide timely feedback on factory information with excessive emissions.

The invention relates to the technical field of internet of things intrusion detection, in particular to a double-queue memory and structural relationship distillation method for intrusion detection, which comprises the following steps of: enabling a model to learn feature representation with distinction degree through a double-queue collaborative memory network under the condition of completely not needing to mark data; according to the method, distance and angle relations in a sample feature space are comprehensively modeled by using a symmetric spatial structure relation distillation method, and a student model of lightweight design efficiently learns structural relation knowledge in the feature space of a teacher network through a teacher-student distribution bidirectional alignment mechanism, and linear fine adjustment is performed on the student model through a small number of labels. The effective intrusion detection system which is lightweight, can process data in real time and can run in an unsupervised mode is obtained, the problem of label dependence is relieved, and the method is suitable for an Internet of Things scene with limited resources.

The application discloses a profile error calculation method for a numerical control machining device, and relates to the technical field of numerical control machining.The method takes a time point corresponding to an expected position point as an initial value of a Newton iteration method, determines a potential time interval in which a time point corresponding to a minimum value of a profile error evaluation function is located by using a gradient descent method, and then dynamically switches the Newton iteration method and a dichotomy method to accurately solve a target time point in the potential time interval, so that the method quickly and reliably converges to a local optimal point closest to the expected position point.The method has the speed of the Newton iteration method and the stability of the dichotomy method, improves the reliability and robustness of a profile error calculation result, and provides an efficient and stable calculation basis for subsequent profile precision control and real-time compensation.

This utility model discloses an audio processing module for an industrial camera with sound field direction analysis, including a microphone array containing at least three asymmetrically distributed directional microphones, configured in the circumferential edge region of the industrial camera housing, for acquiring multi-channel audio signals within the target sound field range; through the setting of high-precision sound field direction, and through a dual-layer asymmetrical microphone array (6 cardioids in the inner ring + 12 supercardioids in the outer ring) and MVDR beamforming algorithm, a 2.5° azimuth resolution is achieved in the 12kHz frequency band, which is 67% higher than that of traditional uniform arrays, and effectively suppresses multipath reflection interference in the industrial environment (positioning error < ±5cm / 3m).