6results about How to "Guaranteed extraction accuracy" patented technology

The application is suitable for the technical field of computer application, and provides a two-dimensional code image feature extraction method and device based on a CNN-MALT hybrid architecture, which comprises the following steps: inputting a target two-dimensional code image into a CNN preprocessing module of a preset lightweight image feature extraction network to generate first structure perception features; inputting the first structure perception features into a MALT encoder of the preset lightweight image feature extraction network, performing feature extraction on the first structure perception features based on a proxy attention mechanism to generate deep features; and inputting the deep features into a feature fusion module of the preset lightweight image feature extraction network to generate target features. Thus, the structure features of the two-dimensional code image are perceived by the CNN preprocessing module, the deep features are generated by the MALT encoder based on the proxy attention mechanism, and the deep features are fused by the feature fusion module, so as to ensure that the calculation complexity is reduced while the two-dimensional code image feature extraction accuracy is improved.

The application relates to the technical field of fluid control and pipeline safety protection, and discloses a fluid leakage automatic response emergency closing valve system based on adaptive control, which comprises a main valve, an actuator, a bypass, a pilot valve, a pressure sensor and a calculation node. The calculation node controls the actuator to drive the main valve to perform a primary cut-off action; the pilot valve is driven to generate a quadrature encoding transient pressure wave and inject the downstream pipeline through the bypass; a transient pressure time sequence signal is acquired, and adaptive wave speed data is output by calling a correction model when phase change of fluid is determined; a net echo signal is extracted based on cepstrum analysis, and the actual physical distance of a leakage point and an equivalent leakage hole cross-sectional area are calculated in combination with the adaptive wave speed data; and then the actuator is controlled to adjust a secondary speed reduction closing trajectory of the main valve. The application can effectively compensate wave speed attenuation caused by phase change to improve ranging accuracy, and adaptively match a fast closing mode or a slow closing mode, so that leakage blocking and water hammer protection are considered.

This invention provides a calibration error compensation method for a high-speed measurement system, relating to the fields of photoelectric non-contact, high-precision, and high-dynamic measurement technology. It solves the problems of error compensation for linear images not being parallel to the vertical direction of the image plane and the modeling of point-line imaging models. First, the centerline of the acquired linear image is fitted. Then, a multi-objective optimization model is established for the centerline, optimizing it into a straight line parallel to the vertical direction of the image plane to compensate for the centerline positioning algorithm error. Next, based on the relationship between the luminescent target and its projection onto the imaging plane as a straight line, multiple rays are established. A spatial straight line parallel to the centerline of the linear image is randomly selected from these rays, and a plane equation is established with the luminescent point. Finally, the plane equations of multiple luminescent targets are solved simultaneously to obtain the calibration parameters. This invention provides a calibration error compensation method for a high-speed measurement system that compensates for identification and positioning errors while simultaneously achieving calibration, with high calibration accuracy and good environmental robustness.

The invention discloses an extraction method for text events with overlapping and nesting characteristics, which comprises the following steps of: performing attention interaction on original text word embedding and event type word embedding through a symmetric bidirectional attention mechanism to obtain corresponding word embedding after semantic enhancement; original text word embedding and semantic enhanced original text word embedding are carried out, important texts are extracted by extracting weights, and the important texts are integrated to obtain optimized text word embedding; splicing the original text word embedding, the optimized text word embedding and the corresponding event type word embedding after semantic enhancement to obtain a spliced text word embedding, and obtaining a trigger according to the spliced text word embedding; and obtaining the start and end probabilities of each word embedded in the spliced text word in the question and answer task, and obtaining the event element of the current question role through a position decoding algorithm to complete the event extraction process. According to the method, overlapped and nested events can be better understood, and event type information in a complex structure can be represented.

The invention relates to the technical field of intelligent medical treatment, and discloses a comprehensive obesity evaluation model based on user health data and application thereof, and the model comprises a data acquisition module, a data preprocessing module, a feature engineering module, a model training module and an evaluation output module. According to the method, multi-source heterogeneous health data are integrated, the limitation of traditional single data assessment is broken through, multi-dimensional factors influencing obesity can be comprehensively captured, data preprocessing and feature engineering optimization are combined, the data and feature quality is effectively improved, and a foundation is laid for accurate assessment; secondly, by adopting a deep learning and causal reasoning hybrid architecture, not only is feature extraction and evaluation precision guaranteed, but also the function logic of the features on obesity evaluation is clarified through interpretability analysis, and the evaluation credibility is improved; besides, the dynamic obesity risk score updated along with the real-time health data of the user is generated, the risk change can be reflected in real time, and accurate and efficient technical support is provided for personalized health management and obesity intervention.

The invention discloses a method and a system for acquiring a low-gray-value light bar sub-pixel edge on the surface of a forging in a large high-temperature forging image, and relates to the technical field of computer vision detection. Comprising the following steps: in a high-temperature forging image, acquiring pixels in a circular area with the diameter of N pixels by taking the pixel-level edge center of a low-gray-value light bar as a central point; based on the gray level of pixels in the circular area and a plurality of pixel-level edges adjacent to the pixel-level edge of the low-gray-value light bar, obtaining the position of a circular area segmentation straight line and a straight line equation; straight lines which pass through the pixel-level edge center of the low-gray-value light bar and are perpendicular to the circular area segmentation straight line are obtained, and the intersection point of the two straight lines is a sub-pixel edge corresponding to the pixel-level edge. The method can effectively solve the problem that the extraction precision of the sub-pixel edge is reduced due to low brightness of the light bar in the high-temperature forging image, and is particularly suitable for accurately calculating the sub-pixel edge of the light bar with a low gray value.