4results about How to "Enhanced Semantic Information" patented technology

This invention discloses a small target detection method and system based on improved YOLOv11 and fused attention, applicable to small target detection in unmanned aerial vehicles (UAVs). The method includes the following steps: constructing a backbone network using a custom module SFBCR; constructing a neck network using a custom module MFBCR to obtain a custom model; and designing a fused attention mechanism based on the SimAM attention module and PartialConv partial convolutional modules. The model parameters are determined by training on the public UAV small target dataset VisDrone. Compared with existing neural networks, the proposed detection method achieves an average accuracy improvement of 2.03% in UAV small target detection applications while maintaining detection speed. It can be used in smart city construction to improve urban traffic management.

The invention provides a structured annotation enhanced vulnerability detection method based on dynamic path selection, and belongs to the field of source code vulnerability detection. Comprising the steps that 1, each code snippet is analyzed into an abstract syntax tree, and a corresponding control flow graph is constructed; 2, evaluating the importance of paths in the control flow diagram, and dynamically selecting a plurality of paths; 3, the code snippets are provided for the large language model, corresponding row-level annotations are generated for each row of codes, and the annotations are inserted into a selected path according to a predefined annotation insertion rule; and step 4, coding the annotation paths into vectors, extracting a local dependency relationship between the paths, modeling a global dependency relationship between the paths, and finally performing vulnerability detection based on fusion features. According to the model provided by the invention, a code structure and a vulnerability situation can be comprehensively understood, and more efficient and more accurate vulnerability detection is realized.

The application discloses a kind of based on deep feature fusion's cable membrane structure digital twin model reconstruction method, including survey entity cable membrane space structure, arrangement acquisition station, acquisition cable membrane space structure three-dimensional point cloud data;The three-dimensional point cloud of cable membrane space structure is preprocessed and constructs dataset;RPMNet network architecture is improved, and spatial attention mechanism method is applied to the point cloud feature extraction layer of RPMNet network architecture;The three-dimensional point cloud of cable membrane space structure is finely registered;Based on the three-dimensional registration point cloud of cable membrane space structure, fine model reconstruction is carried out;Deviation analysis is carried out to the fine model reconstruction of cable membrane space structure and theoretical design model.The method realizes the accurate registration of three-dimensional point cloud data of cable membrane space structure, completes the high-fidelity digital twin model reconstruction of entity cable membrane structure space form, has the characteristics of high efficiency, low cost and wide coverage, has good applicability to cable membrane space structure, and can be widely applied to engineering practice.

This invention provides an image-assisted 3D object detection method, comprising the following steps: Step 1, extracting features from a point cloud containing the 3D object to be detected and its corresponding image; Step 2, fusing the features of the point cloud and the image based on a block self-attention mechanism to obtain fused multimodal features; Step 3, generating a foreground point mask based on the fused multimodal features and using it to extract multi-scale features with a focus on the foreground, obtaining feature maps of the point cloud at different resolutions; Step 4, generating an object detection proposal, i.e., interpolating the feature maps at different resolutions based on the three nearest neighbor interpolation method, fusing the interpolated features with the multimodal features using a multilayer perceptron, and feeding them into a two-stage detection module to generate an object detection proposal, thereby completing the image-assisted 3D object detection.