21results about How to "Reduce false match rate" patented technology

This invention discloses a method and system for UI element localization in cross-platform, full-scenario RPA based on functional semantics, belonging to the field of robotic process automation technology. It includes a pre-construction step, an element recording step, and an element playback step. The pre-construction step involves pre-constructing a UI element functional semantic ontology library specific to the RPA business scenario. This ontology library includes a standardized element functional semantic tag system independent of the client side, business context rules corresponding to each semantic tag, and interactive behavior features. By pre-constructing a client-side independent functional semantic ontology library specific to the RPA business scenario, and using business functional semantics as the core, this invention constructs a cross-platform generalized feature vector for target elements, freeing them from the binding limitations of recording client features. After the same business process is recorded on the Web client, it can be directly run on PC desktop clients, mobile mini-programs, domestic terminals, and closed systems without interfaces without modification, effectively solving the problem of high repetitive development costs in cross-platform scenarios.

The invention provides a color feature extraction and correction method and system for drug sensitive plate MIC value interpretation and a computer readable storage medium. The method comprises the following steps: acquiring a to-be-processed drug sensitive plate image; performing visual preprocessing such as exposure, contrast and white balance on the image according to the configuration file; high-precision alignment correction of the image is realized by using ORB feature detection and a BFMatcher matching algorithm; generating a transparent mask with an Alpha channel based on predefined hole position contour data, and performing non-destructive target segmentation on the aligned image; and converting the target region to an HSV color space, extracting H-channel hue features, executing cyclic correction, eliminating the problem of numerical value jump of the red region, and obtaining stable color feature data. The method has the advantages of being high in image registration precision, accurate in segmentation, stable in color feature, high in robustness and the like, and automatic and standardized interpretation of the MIC value of the drug sensitive plate can be achieved under different shooting conditions.

The application discloses a kind of based on the task cooperation management method and system of offline application, it is related to offline data management technical field, the method includes: determining task cooperation type, and obtaining collaborative task resource package;Task cooperation system environment is generated based on collaborative task resource package;Receive collaborative data, the modification probability of collaborative data is predicted by the prediction model pre-trained in collaborative task resource package, and the collaborative data of high-risk prediction result is early warning and checking;According to the compression packing of collaborative data according to collaborative task type, and when restoring network connection, the packaged collaborative data is executed upload by priority scheduling mechanism.The fusion technical scheme of intelligent matching resource package, generating system environment with edge computing, pre-training model early warning and checking, priority scheduling upload effectively reduces resource package mismatch and parameter error rate, while reducing upload delay, improves offline cooperation efficiency and data reliability.

The application belongs to the technical field of point cloud processing, and discloses a multi-modal semantic mapping method and system based on a master-slave architecture. The method comprises the following steps: S1, extracting feature points in point cloud data and calculating the covariance matrix of the neighborhood point cloud of each feature point; S2, solving the optimal pose corresponding to the minimum of the target function; S3, converting the obtained point cloud data into a world coordinate system using the optimal pose, fusing all frame point cloud data in the world coordinate system to form a three-dimensional point cloud map, projecting the three-dimensional point cloud map to a camera coordinate system, and inputting the three-dimensional point cloud map projected to the camera coordinate system into a semantic segmentation neural network to obtain the semantic label corresponding to each point cloud in the three-dimensional point cloud map, thereby achieving semantic mapping. Through the application, the problems of weak dynamic adaptability, insufficient model generalization ability and long-term running precision decay of the traditional three-dimensional semantic mapping method in a dynamic complex scene are solved.

The application relates to the technical field of building contour extraction, and discloses a building roof contour extraction method and device, electronic equipment and a storage medium, the method comprising the following steps: extracting a building boundary from an optical remote sensing image, and extracting a building vector frame from a vector map; performing global search in an initial search space by using a normalized cross-correlation similarity measure, performing template matching, and preliminarily positioning a building roof contour; calculating a global offset main direction angle, setting an angle tolerance, taking the global offset main direction angle as a central axis, taking the angle tolerance as a half width, generating a sector search space, performing local optimization in the sector search space by using the normalized cross-correlation similarity measure, performing template matching, iteratively adjusting the angle tolerance, converging the sector search space, repeatedly performing template matching, and obtaining a final building roof contour. The application carries out hierarchical matching from global search to local optimization, uses the normalized cross-correlation similarity measure template matching, and realizes building roof contour extraction.

This application relates to the field of image processing technology for fly traps, providing an adaptive stitching method and system for fly trap images. The method includes: extracting feature vectors from multiple images using the SIFT algorithm and a ResNet network respectively, and weighting and fusing the two features using adaptive weight coefficients 'a' generated by the ResNet network to obtain a fused feature vector; based on the fused feature vector, using an improved random sampling consensus algorithm to solve the homography matrix between the two images, performing non-uniform random sampling based on similarity during the sampling stage, and using the top K% of the most similar inliers to form a refined inlier set for calculation during the calculation stage; and completing image stitching based on the homography matrix. This application solves the technical problems of difficult feature matching, stitching misalignment, and obvious fusion traces caused by the small size, dense distribution, and high similarity of insect targets in fly trap images, thus improving the robustness, accuracy, and efficiency of image stitching.

The application discloses a wave image matching method based on fusion matching cost, and relates to the field of stereo matching, and comprises the following steps: S1, left and right eye images of waves are collected by using binocular cameras to determine a disparity range; S2, the left and right eye images of original waves collected in the step S1 are preprocessed; S3, a matching cost value of the wave images is obtained by using an improved Census algorithm; S4, improved Census cost, AD cost and gradient cost are fused to obtain a cost space; S5, a cross-domain is used for cost aggregation; S6, a winner-takes-all (WTA) strategy is used to calculate the disparity of the wave images; S7, the disparity map in the step S6 is subjected to left-right consistency checking, and is subjected to hole filling and sub-pixel optimization processing. The application adopts the above method, solves the problem that the disparity map obtained from the wave images is prone to a large number of invalid points and mismatching points, improves the correctness of wave stereo matching, and guarantees the matching precision of the depth discontinuous region.

The application provides a kind of robust laser SLAM repositioning and mapping method for embodied intelligent robot in the technical field of robot autonomous navigation, comprising: receiving original laser radar point cloud and IMU data, pre-processing point cloud to extract feature points;IMU pre-integration provides initial pose constraint;Loop closure detection is through the two-stage cascade of ScanContext global search and LinK3D local matching, obtains repositioning constraint;Front-end odometry matches feature points with local voxel map, combines IMU optimization to output pose estimation, filters key frame to generate laser odometry factor;With repositioning constraint as initial value, use improved ICP to accurately register to obtain repositioning factor;Finally, construct factor graph, jointly optimize IMU, laser odometry and repositioning factor, solve global consistent pose sequence and map.The application has the advantages that: it greatly improves the system robustness, repositioning accuracy and global map consistency of quadruped robot in severe motion and large-scale scene, while meeting the real-time constraint of embedded platform.

The invention provides a power distribution overhead line three-dimensional modeling method, system and device. The method comprises the steps that an unmanned aerial vehicle is used for collecting an urban area power distribution overhead line image according to a flight path; performing line pixel extraction on the image by using a dual-branch network of odd-even channel splitting to obtain a joint segmentation image; constructing a wire harness direction matching window in the local area based on the joint segmentation map, and generating a line three-dimensional point band through parallax calculation; and tracking the three-dimensional point belt of the line by adopting a progressive search strategy with a preferential maximum curvature radius, and generating a three-dimensional skeleton of the power distribution overhead line. The track of the invention can effectively avoid shielding in a complex urban area. The double-branch network can improve the recognition precision of a fine line target, and effectively isolates multi-directional interference textures such as communication lines and advertising boards. The wiring harness guided matching window fundamentally avoids the introduction of background noise, reduces the mismatching rate and improves the matching calculation efficiency. A smooth and continuous line skeleton can be generated through a progressive search strategy with the priority of the maximum curvature radius.

The invention discloses an oral cavity three-dimensional scanning system and method based on dual-band heterogeneous structure light coding, and belongs to the technical field of optical three-dimensional scanning. The system comprises a light source, a coding light modulation device, at least two image acquisition units and a data processing unit, the light source can respectively emit first wave band light and second wave band light with different wave bands; the coding light modulator modulates the first wave band light into a main coding pattern containing discontinuous stripes and modulates the second wave band light into an auxiliary coding pattern containing continuous stripes, and the stripes of the two coding patterns are alternately arranged at intervals; the modulated first wavelength light and the modulated second wavelength light irradiate the surface of a measured area, and each image acquisition unit acquires a coded image reflected by the measured area; and the data processing unit decodes the main coding image to obtain a reference matching point pair and parallax data, decodes the auxiliary coding image based on the result to obtain a high-density extended matching point pair and extended parallax data, and completes integrated three-dimensional reconstruction through triangulation after integrating the two types of data.

The application discloses a rule intelligent labeling generation method based on task semantic analysis, relates to the technical field of data labeling management, and specifically comprises the following steps: performing zero-sample rule cold start, comprehensively calculating confidence scores according to occurrence frequencies, semantic rationality scores and domain general prior knowledge, and selecting a seed rule set; acquiring the seed rule set, constructing a rule knowledge graph based on multi-dimensional semantic association relationships; acquiring a natural language sentence to be labeled, performing bidirectional refining on sentence semantics and rule semantics of the natural language sentence through a cross-attention mechanism, and calculating a matching score; on the rule knowledge graph, label information is propagated from the seed rule to a candidate rule through an attention mechanism; the propagated label and the matching label are compared, inconsistent rules are marked and preferentially submitted for verification; the rule knowledge graph is updated according to a verification result, and a final labeling rule set is iteratively generated.

The application discloses a dynamic target tracking and ranging method and system based on fusion matching, relates to the technical field of target tracking and identification, and specifically comprises the following steps: collecting video image data of a to-be-measured target and preprocessing; constructing a self-attention enhanced YOLOv8 target detection model to detect the to-be-measured target; predicting a next frame target estimation box of the video image data by using Kalman filtering, updating the target detection box, and positioning the to-be-measured target position; matching the predicted target estimation box with the current target detection box by using the Hungarian algorithm to establish the motion trajectory of the to-be-measured target; converting the coordinates of the to-be-measured target into coordinates in a robot world coordinate system according to the coordinates of the to-be-measured target in a tracking box sequence and based on the internal and external parameters of a camera; and calculating the distance of the to-be-measured target according to the coordinates in the robot world coordinate system. The application realizes accurate target positioning and ranging by converting the target from a pixel coordinate system to a world coordinate system, and solves the problem of inaccurate target distance calculation in a real-time dynamic environment.

The application discloses a complex background infrared image background motion compensation method based on image registration, which comprises the following steps: using a FAST feature point detection algorithm to detect feature points of two adjacent infrared images, and removing redundant feature points by using a quadtree algorithm; generating position descriptors and gray level descriptors, and combining the two with BEBLID descriptors to generate LGB descriptors for feature description of all preprocessed feature points in the two images; dividing each image into four quadrant regions, and calculating hash key values of each feature point; matching feature points in the two images according to the hash key values to obtain matched feature point pairs in the two images; screening the matched feature point pairs by using a random sample consensus algorithm, calculating final background motion parameters, and performing affine transformation on a previous image in the two images according to the final background motion parameters to realize background motion compensation. The application improves the effect of background motion compensation and provides better conditions for subsequent infrared dim and small target detection.

The invention relates to the technical field of information technology, and discloses a scientific research talent and enterprise information matching method and system, and the method comprises the steps: obtaining enterprise demand information, and constructing an enterprise demand feature set; obtaining scientific research achievement information, constructing an achievement feature set, and generating an application scene prediction result; obtaining enterprise historical technology information, constructing an enterprise technology layout feature set, and identifying a potential technology demand direction; obtaining scientific research talent information and constructing a scientific research talent ability feature set; mapping the feature set to a unified feature space to form an enterprise demand vector, an enterprise technology layout vector and a scientific research talent ability vector; and performing matching analysis on the scientific research talent capability vector, the enterprise demand vector and the enterprise technology layout vector based on the unified feature space to generate a matching result set, and outputting matched scientific research talent information and related scientific research achievement recommendation or cooperation docking information. The method can improve the industrialization efficiency of scientific achievements.

The application relates to a SAR image robust registration method based on phase consistency and local scattering entropy, and belongs to the technical field of synthetic aperture radar image processing. The application comprises the following steps: firstly, pre-processing SAR complex data and calculating phase information, extracting stable feature points on the phase image by using the characteristic that the phase consistency is not sensitive to noise; then, calculating the local scattering entropy of each feature point, and constructing a feature descriptor which fuses phase distribution, scattering entropy value and spatial structure information; adopting bidirectional matching and scattering entropy similarity constraint to complete coarse registration and quickly correct large-scale deformation; taking the peak value of scattering entropy similarity in a small search window as a guide, and combining a nonlinear transformation model to realize sub-pixel level fine registration. The application can still stably extract features and realize high-precision matching under strong noise and complex scattering scenes, the registration precision can reach 0.08-0.1 pixels, and the application can be widely applied to the fields of SAR interferometry, ground surface change monitoring, disaster assessment and target positioning.

The application discloses a kind of light cross-view positioning method and system based on quantification knowledge distillation and semantic enhancement matching, comprising: unmanned aerial vehicle aerial image is used as query image, similarity matching is carried out between query image and pre-constructed geographic labeled satellite image database, and the highest similarity matching satellite image is obtained;Respectively extract the texture feature and semantic feature of query image and matching satellite image, and respectively generate corresponding semantic perception descriptor;Based on this, the texture feature similarity matrix and semantic feature similarity matrix between query image and matching satellite image are calculated respectively;Two similarity matrices are multiplied at element level, and the final similarity matrix is obtained, so that the pixel-level key point matching pair between query image and matching satellite image is determined;According to pixel-level key point matching pair and the geographic coordinates corresponding to matching satellite image, the geographic position coordinates of selected point in query image are obtained.The application realizes model light-weight deployment while ensuring high positioning accuracy.

The application discloses a river sudden environmental event early warning intelligent method and system, relates to the technical field of water environment monitoring, environmental emergency and intelligent early warning, and comprises a river network monitoring diagram of upstream and downstream monitoring nodes, risk source nodes and sensitive receptor nodes; online water quality, hydrological and risk source fingerprint data are acquired; upstream and downstream anomaly vectors are generated; a propagation time delay window and an attenuation coefficient are generated according to the length of a river section, flow velocity and flow rate; the upstream anomaly vector is translated in time and corrected in amplitude, and then the propagation consistency is calculated with the downstream anomaly vector; when the propagation consistency reaches a threshold value, only the candidate risk sources that are reachable by a hydrodynamic path are subjected to fingerprint matching, and a graded early warning is output in combination with the expected time window of a pollution group reaching a sensitive receptor and the influence level. The application can reduce single-point false positives, improve the accuracy of sudden water pollution event identification, source tracing and sensitive receptor early warning.

The invention provides a dynamic tracking method, system and device for an explosion fragment and a storage medium, and belongs to the field of optical tests.The method comprises the steps that a Bouguet epipolar correction algorithm is used for processing a binocular high-speed image, and matching search is reduced from two dimensions to one dimension; thirdly, providing a Rank-Census mixed binocular matching algorithm, and realizing high-precision matching and three-dimensional reconstruction of fragments with the same name in a complex environment by fusing the statistical robustness of Rank transformation and the texture sensitivity of Census transformation; secondly, a fragment diffusion physical model is introduced, the explosion center distance and the diffusion angle range of fragments are defined according to the charging structure, and a candidate space area of the fragments at the next moment is predicted; and finally, in the prediction area, adopting a speed consistency screening strategy, calculating an arithmetic mean value of the instantaneous speed of the candidate fragments, and selecting the fragments with the speed closest to the mean value for track association, thereby realizing stable and continuous tracking of the target fragments.

The embodiment of the application discloses a kind of data center's alarm matching method, comprising: obtaining the alarm text of data center, and standardization processing is carried out to obtain standardization alarm text;From the standardization alarm text, extract alarm element features, and carry out different weight processing to form structured alarm feature set;The structured alarm feature set is matched with historical alarm library to determine corresponding historical alarm record;Based on the historical alarm record, extract alarm disposal information, and execute adaptive disposal result according to the alarm disposal information.The application scheme forms structured feature set by standardizing data center alarm text, extracting alarm element features and carrying out different weight processing, and then matches with historical alarm library and executes adaptive disposal, solves the defects of low accuracy, poor adaptability and low efficiency of existing alarm matching, supports the stable operation of data center dynamic system equipment and complex alarm scene.

The application discloses a kind of external key point matching and tracking trajectory calculation method and device based on VGGT.The method comprises: loading configuration file and initializing model;Load external key point coordinates and carry out standardization and quantity alignment;First, input image is converted into tensor format, then image features are extracted, and finally L2 normalization is carried out;Depth feature vector of external key point is extracted from feature map;Calculate cosine similarity to evaluate matching quality;The average amplitude and standard deviation of feature vector are calculated to evaluate feature quality, and the matching threshold is dynamically adjusted;The coordinate difference of matching point is calculated to eliminate the matching that does not meet geometric constraint, and tracking trajectory is generated;Tracking quality analysis and visualization are carried out.The matching success rate of the application is greatly improved, the false matching rate is obviously reduced, the time consumption of processing image is obviously shortened, the efficiency is also greatly improved, the tracking stability is obviously improved under dynamic interference, and automatic high-precision inspection without manual intervention is realized.

The present application relates to the technical field of information matching, in particular to a past matching method and system based on multi-feature fusion natural language processing, which comprises the following steps: obtaining past description texts from two users respectively, calculating the scores of the two texts in five dimensions of semantic similarity, role complementary similarity, time entity similarity, place entity similarity and name gender similarity respectively, obtaining the comprehensive matching degree through weighted fusion, and completing matching determination and notification interaction based on the preset matching threshold. Through multi-dimensional feature fusion, especially the innovative role complementary similarity calculation mechanism, the accuracy and robustness of the past matching are greatly improved, without the need for the user to provide the accurate identity information of the other party, and the matching can be completed only through the common past description, which can be widely applied to the friend matching scene of social network platforms.