8results about How to "Data processing is simple" patented technology

This application provides a model training method, apparatus, electronic device, computer-readable storage medium, and computer program product. The method includes: extracting audio feature sequences from raw audio data; determining a first phoneme feature sequence based on the raw audio data and text data; aligning the first phoneme feature sequence based on the audio feature sequence to obtain a second phoneme feature sequence; configuring an omission flag in the second phoneme feature sequence based on text data to obtain a third phoneme feature sequence; structurally associating the audio feature sequence and the third phoneme feature sequence to obtain structured data; storing the structured data in a target storage file; and upon receiving a training instruction, reading the structured data from the target storage file and updating the model parameters of the audio synthesis model based on the data in the structured data that does not have an omission flag configured, thereby obtaining the trained audio synthesis model. This application can improve the training efficiency and accuracy of audio synthesis models.

The invention discloses a TDI camera-based product detection method and system, and a medium. The method comprises the following steps: installing a line light source; the light intensity of the line light source is controlled to be in periodical light and shade stroboflash according to stripe distribution, so that stripe images meeting the phase shift method under the RGB three channels are obtained; analyzing stripe images under three channels of RGB, and calculating and obtaining original height data of all pixel positions under each channel of RGB based on phase deflection; traversing the original height data of all pixel positions so as to obtain final height data of the current pixel position through weighted averaging according to RGB channel weights corresponding to the current pixel position; according to the method, the TDI true color camera is used for collecting the three-channel fringe images, channel crosstalk and response inconsistent errors are restrained through a self-adaptive weighted average mechanism, high-speed measurement is guaranteed, meanwhile, the reconstruction precision and robustness of phase deflection are remarkably improved, and false data introduced by interpolation are avoided.

This invention provides a force sensor comprising an annular housing, a top cover and a base respectively fixed to one end of the housing, a loading platform fixed to the top cover near the base and spaced apart from the housing and the base, at least three force-measuring beam assemblies spaced apart from each other around the periphery of the loading platform, and a first sensor device and a second sensor device disposed on each force-measuring beam assembly. Each force-measuring beam assembly includes a horizontal force-measuring beam fixed to the periphery of the loading platform and a vertical force-measuring beam fixed to the loading platform near the base. The horizontal and vertical force-measuring beams in each force-measuring beam assembly are perpendicular to each other and located in the same plane. All force-measuring beam assemblies are rotationally symmetrically distributed. The force sensor of this invention can improve its decoupling effect and simplify data processing during measurement.

The application discloses a network communication method and device, electronic equipment and medium, and relates to the technical field of computers. The method is applied to a first core in a heterogeneous multi-core processor. The heterogeneous multi-core processor further comprises a network controller connected with the first core. The method comprises the following steps: in the case that the network controller receives first network data, the first network data is parsed to obtain a network port corresponding to the first network data; based on the network port, a target core corresponding to the first network data in the heterogeneous multi-core processor is determined; in the case that the target core is a second core, the first network data is stored in a preset shared memory corresponding to the second core, and a first interrupt signal is generated. Through the application, each core in the heterogeneous multi-core processor can communicate with the outside through one network controller.

This invention relates to the field of ground testing technology for spacecraft attitude control systems, specifically providing a ground testing method for the accuracy of a multi-level composite attitude control system. First, a fully physical simulation ground testing platform is constructed, including an on-orbit environment simulation system and a multi-level composite attitude control simulation system. Two reference calibration devices are built using a calibrated dual-axis photoelectric autocollimator, and two measurement reference units are constructed using the dual-axis photoelectric autocollimator and a multifaceted prism. Based on a cubic mirror, the calibrated dual-axis photoelectric autocollimator, and the dual-axis photoelectric autocollimator, zero-position reference calibration and angle calibration are performed on the multi-level composite attitude control simulation system. After calibration, the multi-level composite attitude control simulation system is controlled to move from the initial position to the final position, and angle information is collected through the two measurement reference units. The attitude control accuracy is calculated based on the angle information. This invention realizes ground testing of a multi-level composite attitude control system based on fully physical simulation.

The invention provides a gene expression profile bi-clustering method based on a swarm intelligence algorithm, and belongs to the technical field of gene expression, and the method comprises the following steps: constructing an index matrix for a gene expression data matrix, obtaining a bi-clustering seed, initializing a population, taking the bi-clustering seed as a first population individual, and constructing a second population individual; adding the first population individuals into a second population for population expansion to obtain two population individuals, evaluating the first population individuals by using a fitness function, and evaluating a plurality of individuals which contribute to the second population to the greatest extent by the individuals in the first population until a preset maximum number of evolution times is reached; and selecting an optimal bi-cluster from the second population. The index matrix is constructed for the gene expression data matrix, so that later data processing is simpler and higher in efficiency, meanwhile, data values of each gene under each condition can be searched by using the swarm intelligence algorithm, and therefore, the search range is wide, and important biological information is not easy to miss.

The invention provides a pipeline defect identification sensor, device and method wherein the pipeline defect identification sensor comprises a magnet yoke, a high-frequency alternating current coil, a direct current coil and an induction coil; the magnet yoke comprises a cross beam and a side arm, and the cross beam is arranged on the side arm to form an open end; the high-frequency alternating-current coil is sleeved on a cross beam of the magnet yoke so as to apply sine alternating-current excitation and generate a detection magnetic field; the direct-current coil is sleeved on the side arm of the magnet yoke to apply direct-current excitation, so that the detection magnetic field is saturated or nearly saturated; the open end of the magnet yoke is located above the pipeline, so that the pipeline generates an eddy current field based on the saturated or near-saturated detection magnetic field; and the induction coil is arranged between the magnet yoke and the pipeline to induce an eddy current field generated by the pipeline to obtain an induction signal so as to detect pipeline defects.

The application discloses an online water quality analysis method and system based on a multispectral image analysis model, and belongs to the technical field of water quality analysis, and comprises the following steps: collecting a multispectral image of a target water body and performing pretreatment, and constructing a multispectral image analysis model comprising an analysis model, a classification model, a water body plankton coverage inversion model, a water body transparency inversion model, a water surface eutrophication inversion model and a key water quality parameter analysis model. The analysis model is used to extract water quality characteristics, the classification model is used to extract a water surface line, and then current water quality data and the water surface line are input into the inversion model to perform water body plankton coverage, water body transparency and water surface eutrophication inversion and key water quality parameter analysis, and finally, analysis results are fused to output comprehensive water quality evaluation. The online water quality analysis method provided by the application can realize efficient and real-time water quality monitoring and analysis, and provides a scientific basis for water environment management.