309results about How to "Improve processing precision" patented technology

The invention discloses a method for correcting dialogue intention information of old people based on a deep neural network, and particularly relates to the technical field of dialogue optimization, and the method comprises the steps: collecting and cleaning multi-topic multi-round dialogue data, marking multi-round dialogue corpora, and unifying the format; performing feature extraction on the labeled corpus by using the multi-intention recognition sub-model to generate a multi-intention representation vector capable of representing each intention; screening context content related to the target intention in combination with a hierarchical attention mechanism of user context dynamic weight adjustment; processing context content related to the target intention through gating operation based on multiple rounds of interaction history, and reserving semantic fragments with reference value; performing parallel analysis on the reference semantic fragments based on a multi-task learning structure of a deep neural network, performing classification processing on different intentions, and calculating scores; and executing multi-intention fusion correction operation according to classification and scoring results, and outputting intention information capable of distinguishing each intention.

The invention discloses a medical image data processing method and system, and relates to medical image data processing, and the method comprises the steps: introducing an attention mechanism to calculate the attention weight of each mode based on the specific features of each mode, generating a specific feature map according to the attention weight, and carrying out the recognition of the specific feature map through a feature alignment algorithm, performing spatial alignment on the specific feature maps of the different modalities to generate an aligned specific feature map, and identifying correlation among the modalities by using a multi-head attention mechanism to generate a comprehensive feature map; and based on the comprehensive feature map, carrying out pixel classification through UNet to generate a preliminary segmentation result of the lesion area, meanwhile, utilizing an edge monitoring algorithm to obtain a multi-scale edge feature map of the lesion area, comparing the multi-scale edge feature map with the preliminary segmentation result, removing a pseudo lesion area according to a comparison result, and finally generating a lesion positioning map. According to the method, the pseudo focus area is eliminated through multi-scale edge feature comparison, and the accuracy and robustness of focus positioning are ensured.

The invention discloses a network information age optimization method and system based on meta-deep reinforcement learning. The method comprises the following steps: establishing a video transmission model, and collecting current state data, initial reward information and action information; the method comprises the following steps: acquiring meta-features through a meta-feature extraction function, acquiring a next-step action by an Actor strategy network in combination with a current strategy and the meta-features, sending the next-step action back to a video transmission model to give next-step state data and reward information, evaluating a next-step action effect selected by the Actor strategy network through a Critic value network, and calculating to obtain an Actor strategy network strategy gradient. Meanwhile, updating a value function of the Critic value network, setting a shearing probability ratio mechanism, and performing strategy gradient shearing on the network; and dynamically updating network parameters in combination with a meta-learning algorithm. According to the invention, the overall utilization rate of network resources is improved, and data transmission failure and retransmission caused by network congestion are reduced.

The invention relates to the technical field of financial data processing, in particular to a data processing method and system based on financial information bills. The data processing method comprises the following steps: acquiring multi-modal financial information bill data; and performing bill modal decomposition on the multi-modal financial information bill data to generate financial information bill modal decomposition data. Bond information price income prediction needs a large amount of financial data, such as market interest rate, bond expiration date, face interest rate and the like, the data are similar to the data form of bill modal decomposition, and a more accurate model can be obtained through standardization and feature extraction of multi-modal data. Therefore, valuable input is provided for bond price and income prediction. According to the invention, through multi-modal data fusion, counterfeited feature comparison modeling, abnormal bill clustering analysis and dynamic risk assessment, the security and accuracy of financial bill identification are improved.

The invention discloses a method for analyzing mineral trace elements by utilizing a micro-Raman spectrum, which comprises the following steps of: fixing a mineral sample on a rotary sample table of a micro-Raman spectrometer, selecting an area to be detected by combining a micro-imaging technology, collecting a Raman background signal under the pre-illumination of low-power laser, and analyzing the trace elements in the area to be detected according to the Raman background signal. The laser focal length, the spot size and the incident angle are adjusted accordingly, the laser is focused on a target area, under the calibrated laser condition, through a multi-angle excitation micro-area, after an algorithm is applied to optimize signals, reduce impurity interference and integrate multi-angle Raman signals, the characteristic spectrum information of the microelements can be extracted, and the microelements can be accurately detected. High-resolution analysis of sample components is realized, and a three-dimensional distribution model of microelements in a microcell is constructed in combination with microscopic imaging data. The method has the characteristics of non-destructiveness, high resolution and rapid detection, and is suitable for mineral component analysis in the fields of geology, material science and the like.

The invention discloses an intelligent interactive Agent system based on Langchain and an implementation method, and relates to the technical field of natural language processing and artificial intelligence. Comprising the steps of 1, establishing an intelligent interactive Agent system by using an interface and an integrated tool provided by a Langchain framework, 2, managing calling of various tool modules and services through a central scheduling module, and dynamically registering and unbinding the tool modules by using the interface provided by the Langchain framework, and 3, processing user input by using a large language model through a process management module, and establishing an intelligent interactive Agent system. 4, a series of tool modules are integrated through the integrated tool management module, the tool modules comprise a text search tool, a weather query tool and a video information acquisition tool, each tool module works independently or is in linkage with other tool modules, and the tool modules are connected with the integrated tool management module. And 5, managing a user interaction interface through a front-end management module, supporting a user to query and control in a natural language, displaying a processing progress and a processing result in real time, and providing feedback and help information.

The invention relates to the technical field of television image communication, in particular to a user information input method and system in an intelligent television all-in-one machine. The method comprises the following steps: collecting a user image, carrying out white balance correction and focusing processing, generating a moving object focusing image, carrying out identification and attention tendency analysis on a character in the focusing image, obtaining character tendency data, analyzing a historical interaction image, identifying an interaction mode and constructing a user intention prediction model, and obtaining a user intention prediction result. And when the tendency data shows that the user preferences the television content, intention drilling and auxiliary adjustment are performed on the current image based on the prediction model, predicted user input information is generated, logic conflicts in the predicted information are detected, feedback optimization is performed, and final user input information is output. According to the invention, a more accurate and more flexible user information input method is realized.

The invention discloses a virtual scene construction method based on point cloud data, a medium and equipment. The method comprises the following steps: constructing an initial virtual scene based on original point cloud data; constructing a depth map corresponding to the point cloud data in the initial virtual scene, and determining a shielding relationship among the objects in the initial virtual scene based on the depth map; performing physics-based rendering on an object in the initial virtual scene, and processing the rendered initial virtual scene by adopting global illumination; and integrating the occlusion relationship into the rendered initial virtual scene to generate a new virtual scene. Through the above mode, the method can improve the trueness of building the virtual scene through the radar point cloud data.

The invention discloses a metal surface defect detection and classification method, system, equipment and medium, and the method comprises the steps: processing images of different defect types through gray normalization, adaptive histogram equalization, a Gabor filter, a maximum entropy algorithm, morphological processing and an improved local binary pattern; carrying out defect position and size calculation on the processed image, and carrying out GUI (Graphical User Interface) measurement page design; the system, the equipment and the medium are used for realizing the metal surface defect detection and classification method. According to the method, the image processing effect is remarkable, the image contrast is improved, the defect boundary identification is enhanced, the accurate segmentation of the image and the detailed optimization of the structure are realized, the image features are accurately extracted, and efficient classification training is supported; the method has the advantages of improving image quality, enhancing defect detection precision and optimizing feature extraction and analysis, and can be widely applied to the field of high-precision image analysis such as metal defect detection, industrial quality control and computer vision.

The invention relates to the technical field of signal encryption, in particular to an aviation data bus signal encryption method and system based on a field programmable gate array (FPGA), which comprises the following steps of: based on input three-dimensional tensor signal data, extracting slices from a time index of an aviation data bus according to a fixed interval, progressively sequencing amplitudes of signal channels, and then segmenting; according to the method, through processing logic of matrix decomposition and feature curve generation, the feature extraction and dimension reduction process is optimized, the distinction degree and the expression capacity of feature information are improved, the randomness and the anti-cracking capacity of a key are enhanced through a key generation mode combining a dynamic disturbance factor and an encryption parameter, and the safety of signal transmission is guaranteed; through combination of amplitude superposition and frame-by-frame storage, dynamism and high efficiency of a signal encryption process are realized, signal tampering and eavesdropping risks are reduced, and through feature reconstruction and point-by-point matching processing in a decryption stage, accurate restoration of signal amplitude is ensured, and data integrity and anti-interference capability are guaranteed.

The invention relates to the technical field of data mining, in particular to a land utilization data set processing method and system. The method comprises the following steps: acquiring ground sensing data of a monitoring area, and constructing coordinate system data of the monitoring area; fusing the ground sensing data of the monitoring area to obtain a monitoring area sensing network; acquiring remote sensing data of the monitoring area and performing calculation to obtain land vegetation coverage rate data of the monitoring area; calculating based on the monitoring area sensing network to obtain monitoring area land vegetation coverage rate error data; according to the monitoring area land vegetation coverage rate data and the monitoring area land vegetation coverage rate error data, carrying out influence factor estimation to obtain a resolution influence factor; and carrying out regional land vegetation coverage rate correction on the monitored regional land vegetation coverage rate data to obtain standardized regional land vegetation coverage rate data, and analyzing the standardized regional land vegetation coverage rate data to obtain regional land vegetation coverage rate trend data. According to the invention, the processing precision of land utilization data is improved.

The embodiment of the invention discloses a new infectious disease monitoring and early warning method and device based on multi-point triggering. The method comprises the following steps: acquiring clinical data of a patient in HIS, LIS, PACS, operation and electronic medical records from a medical institution; setting a screening strategy of new infectious diseases; identifying suspected cases according to the clinical data of the patient; identifying abnormal diagnosis and treatment behaviors of the patient from clinical data of the patient based on big data analysis; determining a suspected case level according to the suspected case and the abnormal diagnosis and treatment behavior; obtaining relevant survey data of suspected cases; judging whether the suspected cases have aggregated cases or not according to the related survey data of the suspected cases; and if the suspected cases have aggregated cases, determining an infectious disease transmission route, and performing case management. By implementing the method provided by the embodiment of the invention, real-time monitoring and early warning of infectious disease epidemic situations and other public health events can be realized, and the processing efficiency and accuracy of new infectious diseases are improved.

The invention relates to the field of image processing, and discloses a violation image detection method and device, equipment and a storage medium. The violation image detection method comprises the following steps: acquiring pre-annotation data obtained by annotating a sample image based on a preset model auxiliary annotation system; training an image processing model of a preset decoupling framework based on the pre-annotated data to construct a violation image detection model; and performing violation detection on a target image based on the violation image detection model, the target image being obtained by preprocessing an original image based on a preset redundancy removal algorithm. By using the preset model auxiliary labeling system, the preset decoupling framework and the preset redundancy removal algorithm, the positioning and classification efficiency of labeling and the model training efficiency can be remarkably improved, the model deployment cost is reduced, the model processing efficiency is improved, efficient violation detection of the image is realized, and the detection accuracy is improved. The risk control scene of the vertical screen picture of the mobile phone can be effectively compatible.

The invention relates to a red-black image recognition and extraction method and system based on a power distribution network line, and belongs to the technical field of data processing. The method comprises the following steps: acquiring a red image and a black image of a power distribution network line from a GIS system; traversing the red image and the black image based on a set traversal algorithm, identifying switch equipment nodes in the power distribution network, and constructing a topological structure of the power distribution network; wherein a traversal algorithm is set to express the power distribution network as a graph, then node traversal is carried out, and verification is carried out through a preset rule; and based on the power distribution network topological structure, automatically constructing a power distribution network sketch. According to the embodiment of the invention, a complex and disordered power distribution network red-black diagram is converted into a power distribution network sketch which is matched with the switch equipment nodes more conveniently and visually, a solid foundation can be laid for subsequent intelligent monitoring of the power distribution network, and the data processing precision is improved.

The invention relates to an optical cable route detection method and device, electronic equipment and a storage medium, and belongs to the technical field of optical fiber sensing and sensing networks. According to the method, vibration with specific frequency is emitted, the vibration and irrelevant vibration are distinguished, scattered light in a to-be-detected optical cable is collected to obtain a vibration signal, auxiliary signals are superposed in the scattered light when the optical power is low, the small signal processing precision is improved, and scattered signals are obtained according to the scattered light and processed. Converting the processed scattering signal to a frequency domain according to the vibration duration, accumulating time domain information to improve the frequency domain signal-to-noise ratio, improving the routing judgment accuracy, obtaining the peak-to-average ratio of the frequency domain signal, judging whether the corresponding vibration signal is an effective signal or not, taking the corresponding ground position point as an effective position point, and combining the peak-to-average ratio to judge whether the corresponding vibration signal is the effective signal or not. And obtaining a route position point right above the to-be-tested optical cable, and obtaining the route of the to-be-tested optical cable. According to the invention, the accuracy of vibration detection is improved, so that the accuracy of optical cable routing detection is improved.

The invention discloses a welding quality on-line detection method based on a laser welding machine, and relates to the technical field of welding process quality detection.The welding quality on-line detection method comprises the steps that a welding seam image in the welding process of the laser welding machine is collected, the welding seam image is converted into a gray level image, and binarization processing is conducted on the gray level image; performing morphological processing on the binarized image, eliminating noisy points in the image through opening operation and closing operation, and extracting edge contour features of the weld seam; detecting a straight line segment in the weld edge contour features, and calculating geometric parameters of the straight line segment; and analyzing the geometric parameters by adopting a quality scoring model to generate a weld quality score value. Through improved image processing, an efficient and accurate solution is provided for weld joint quality evaluation in the laser welding process, the accuracy of weld joint quality detection is improved, the real-time performance and the intelligent level of the evaluation process are optimized, and the method has wide application prospects.

The invention relates to the technical field of video processing, and discloses a cloud video processing method based on a truck driving recorder, which comprises the steps of video data uploading, video feature extraction, video feature analysis and warning signal display. Through complex calculation and analysis models, the fatigue degree of the driver can be accurately evaluated, and the accuracy and reliability of driving analysis are improved; multiple factors such as the continuous driving duration of the driver, the eye closing rate of different time periods, the eye closing judgment factor and the abnormal driving coefficient are comprehensively considered, the state of the driver is comprehensively evaluated, and the limitation of single-index evaluation is avoided; the warning signal is visually displayed to related personnel, so that the related personnel can quickly take measures, such as reminding a driver of having a rest or performing other interventions, and the safety of the driver and road traffic is guaranteed.

The invention relates to an automatic shot blasting equipment system for complex parts based on intelligent identification, which mainly comprises three parts, namely shot blasting equipment, a control device and a machine vision identification module, a workpiece mechanical arm capable of grabbing a workpiece and a spray head mechanical arm capable of grabbing a spray head are arranged in a treatment bin of the shot blasting equipment; the workpiece mechanical arm is mounted at the top of the treatment bin, and the spray head mechanical arm is mounted on the side wall of the treatment bin; the control device is used for controlling the shot blasting equipment to work, the machine vision recognition module is used for scanning to obtain workpiece image data and transmitting the workpiece image data to the control device, and the control device analyzes the position and posture of each related part in a workpiece according to the workpiece image data and is used for controlling the shot blasting equipment in real time. The device is simple in structure, capable of implementing automatic control and suitable for complex workpieces.

The invention relates to the technical field of power plant inspection, in particular to an information processing method and device of an inspection robot, equipment and a storage medium, and the method comprises the steps: collecting an equipment image in a moving inspection process of the inspection robot, carrying out the instrument detection based on the collected equipment image, and determining an instrument region image in the collected equipment image; the method comprises the following steps: reducing noise influence in the environment, correcting and identifying an instrument area image, extracting instrument information of each equipment image acquired in the inspection process, improving the information processing precision of an inspection robot in the inspection process, and finally performing equipment self-inspection according to the equipment instrument information, thereby improving the safe operation capability of equipment in a power plant. The technical problem that in the prior art, the inspection efficiency of equipment in a power plant is low is solved.

The invention relates to a data processing method and device based on a hybrid expert model, computer equipment, a computer readable storage medium and a computer program product. The method comprises: based on a target character in input data, determining a target probability corresponding to the target character, and determining a target probability difference threshold corresponding to the target character, the target probability comprising a probability estimation value of each expert network model in a preset hybrid expert model; screening the expert network models based on the probability estimation values of the expert network models and a target probability difference threshold to obtain a target expert network model; and processing the target character based on each target expert network model to obtain a data processing result. By adopting the method, the models contained in the expert model can be dynamically mixed based on the probability difference threshold, the data processing of the character dimension of the input data is realized, and the adaptation degree between each dynamically determined expert network model and the characters in the input data is ensured.

The invention discloses an audio frame insertion method and device, equipment, a medium and a program product. The method comprises the steps of obtaining audio data; wherein the audio data comprises a plurality of audio frames; recognizing lost audio frame information based on the multiple audio frames through a preset audio frame recognition model; wherein the lost audio frame information comprises the position and the number of audio frames; predicting a to-be-inserted audio frame based on the multi-frame audio frame and the lost audio frame information through a preset prediction model; and performing frame insertion processing on the audio frame to be inserted according to the lost audio frame information. By adopting the method provided by the embodiment of the invention, the frame insertion processing speed and precision can be improved, and the problem of sound and picture desynchrony during video playing is effectively solved.

The invention provides a gear ring rotational separation machining method and equipment. The equipment comprises a workbench, a workpiece sliding table, a tool sliding table, a workpiece main shaft assembly and a tool main shaft assembly. The workpiece sliding table and the tool sliding table are installed on the workbench, the workpiece sliding table can move transversely, and the tool sliding table can move longitudinally. The workpiece main shaft assembly is installed on the workpiece sliding table, and the cutter main shaft assembly is installed on the cutter sliding table. The workpiece main shaft assembly is used for being installed on a workpiece body, and the tool main shaft assembly is used for installing a tool body. The machining precision is improved, and the structural organization of a workpiece is not damaged; and workpieces are simply and conveniently machined, blades are easy to replace, stability is high, and machining efficiency is high.

The invention relates to a semi-supervised underwater sound event detection method based on joint disturbance consistency constraint. The method comprises the following steps: acquiring logarithmic Mel spectrums of strong label audio data, weak label audio data and unlabeled audio data; performing weak enhancement and strong enhancement processing on the logarithmic Mel spectrum of the unlabeled audio data to obtain a corresponding enhanced spectrum; constructing a deep learning network model, and performing supervised training by using the frequency spectrums of the strong label audio data and the weak label audio data; adopting a multi-feature disturbance and strong disturbance combined semi-supervised model for the unlabeled audio data; and finally, optimizing the training process of the deep learning network model by calculating a joint loss function. And detecting an underwater sound event by using the trained deep learning network model, and outputting an event category of the audio data. The method is used for realizing effective detection of the sound event by using a small amount of strong injection data and improving the model performance.

The invention discloses an alarm clustering topology method and system based on multi-dimensional features, belongs to the technical field of network security, and aims to solve the technical problems of poor timeliness, multiple repeated false alarms and low association degree of network alarm information processing. Comprising the steps of collecting original alarm information; performing multi-dimensional aggregation on the pre-processed alarm information based on a set alarm dimension; calculating the alarm weight of each alarm set based on the weight of each reference factor; clustering each alarm set based on the new alarm dimension; screening the alarm information of alarm convergence based on a predefined rule base; calculating the alarm number and the alarm weight sum between each pair of IP node addresses, and calculating the thickness and the color depth of a side line according to the normalized weight sum; an IP node topological graph and side boundary thickness and color depth information are constructed through a visual tool, and a topological relation graph between IP nodes is drawn; and visually displaying the alarm information.

The invention relates to the technical field of data analysis, and discloses a slurry pump fault detection and service life prediction system and method based on big data. According to the method, firstly, a slurry pump vibration signal is obtained, an abnormal vibration signal is detected and deleted, and then noise reduction processing is conducted on the slurry pump vibration signal; secondly, after the denoised slurry pump vibration signal is subjected to segmented aggregation approximation and Grubby angle field processing, a slurry pump vibration signal image is generated, the characteristics of the slurry pump vibration signal image are extracted, a convolutional neural network is trained, convolutional neural network parameters are optimized based on an electric eel foraging optimization algorithm, a convolutional neural network recognition model is obtained, and a recognition result of the convolutional neural network is obtained; fault detection of the slurry pump is realized; and decomposing the IMF component of the slurry pump vibration signal, training the LSTM neural network to obtain an LSTM neural network prediction model, and outputting a slurry pump life prediction value. The purposes of fault detection and service life prediction are achieved by processing the vibration signal of the slurry pump, and the method is accurate and objective.

The invention discloses a bridge point cloud extraction method based on evolution gradient search, and relates to the technical field of data processing, and the method comprises the steps: employing a clustering algorithm to carry out the denoising of a bridge surface point cloud; preliminarily extracting the de-noised bridge floor point cloud based on a multi-thread and minimum heap data structure, and meanwhile, setting a perpendicularity density constraint to further extract the bridge floor point cloud; a gradient-based optimization algorithm is combined with an evolutionary strategy to serve as a complementary algorithm to be incorporated into a unified framework, and the extracted bridge deck point cloud is optimized by alternately adopting a gradient updating method and an evolutionary updating method, so that the finally extracted bridge deck point cloud is obtained. According to the method, the processing precision of the bridge deck point cloud is improved, the main body structure of the bridge can be more effectively identified and separated from the complex point cloud data by introducing the evolution gradient search strategy, and meanwhile, the interference of noise and redundant information on a final result is remarkably reduced.

The invention relates to the technical field of battery electronic device recovery, in particular to waste battery electronic device recovery processing equipment which comprises a processing chamber, a conveying mechanism is arranged on the upper portion in the processing chamber, a material inlet is formed in the upper portion of one side of the processing chamber, and a material output port is formed in the lower portion of the other side of the processing chamber. A frozen gas tank, a low-temperature cooling box and a crusher are sequentially arranged below the conveying mechanism in the treatment chamber from left to right, and a powder guide-out opening extending to the material output opening is formed in the lower portion of one side of the crusher; a lifting hoisting mechanism is installed on the conveying mechanism, a containing basket used for containing waste battery electronic devices is arranged on the hoisting mechanism, the outer wall of the containing basket is composed of a plurality of stop levers arranged at intervals, and an electric door is arranged at the bottom of the containing basket. The device has the advantages of being simple and compact in structure, capable of improving personal safety, more energy-saving and more environment-friendly.

The invention relates to the technical field of intelligent mine mining, in particular to an intelligent mine physical property parameter real-time inversion and coal mining simulation real-time optimization method. The method specifically comprises the steps of arranging monitoring points, and monitoring physical property parameter stress data sigma and displacement data u in real time; establishing a mathematical model composed of a geometric equation, an equilibrium equation, a constitutive equation and initial-boundary conditions; a solver is used for simulating the coal mining process, and simulated stress data sigma simulation and displacement data u simulation of each monitoring point are obtained; comparing the real-time simulation stress data sigma simulation and the displacement data u simulation of each monitoring point with the field real-time monitoring stress data sigma monitoring and the displacement data u monitoring, and constructing a loss function; and by utilizing an optimization algorithm, the physical property parameters of the stratum are determined while the coal mining simulation is more accurate. According to the method, the accuracy and optimization effect of coal mining simulation can be improved, so that the safety and resource utilization rate of mine mining are improved.

The invention relates to the technical field of signal processing, and discloses a harbor equipment vibration signal processing method, which comprises the following steps: constructing a comprehensive function, and calculating a harbor equipment vibration signal based on the characteristics of a frequency scale, a second wavelet basis function and a threshold value; the concentration ratio of energy distribution, the calculation complexity of second wavelet basis functions under the frequency scale, the orthogonal value between the second wavelet basis functions under the frequency scale, the sparse value of second wavelet coefficients under the frequency scale and the signal-to-noise ratio of time domain signals reconstructed after threshold processing are comprehensively considered. The selection of the optimized frequency scale, the optimized second wavelet basis function and the optimized threshold is realized. The innovative method gets rid of the limitation of a single or optimized algorithm, fixed parameter setting or predefined rules in the prior art, and significantly improves the precision and robustness of low-frequency vibration signal processing.

The invention provides a measuring point selection method and system based on a 3D radar level gauge, and relates to the technical field of measurement, and the method comprises the steps: obtaining original 3D radar point cloud data of a measurement scene, carrying out the adaptive space division of the measurement scene through employing an improved octree algorithm, obtaining a three-dimensional reconstruction result of the measurement scene, and carrying out the calculation of the three-dimensional reconstruction result. Performing normal vector calculation and region clustering, and determining a plurality of key regions in the three-dimensional reconstruction result in combination with a region growing and merging algorithm; and taking maximization of the coverage degree of the measurement points to the key area as a first target, taking minimization of the geometric distribution difference degree between the measurement points as a second target, solving the single-target optimization problem through an improved differential evolution algorithm, and selecting a solution with an optimal comprehensive evaluation index as a final 3D radar level gauge measurement point optimization arrangement scheme.