78 results about "Visual placement" patented technology

The invention belongs to the technical field of cyberspace surveying and mapping, and particularly relates to a visual position identification method. According to the method, a trained global feature extraction network is utilized to perform feature extraction on a to-be-queried image and each reference image in a reference library to obtain respective global descriptors, and the similarity between the to-be-queried image and each reference image is calculated according to the respective global descriptors. Identifying the position of the to-be-queried image according to the geographic position of the reference image with the highest similarity; wherein the global feature extraction network firstly extracts features of different scales of an input image, and fuses the features of different scales to obtain a fused feature; carrying out feature enhancement and redundant noise suppression processing on the fusion features by utilizing relation perception global attention to obtain enhanced features; and finally, carrying out feature coding on the enhanced features by utilizing MixVPR so as to generate global descriptors. According to the method, the effect of integrally improving the robustness and generalization in coping with factors such as image view angle and complex appearance environment change and the like can be achieved.

The invention discloses a webpage data collection method and a related device, and relates to the technical field of data processing.The webpage data collection method comprises the steps that a webpage data collection knowledge base containing data collection modes of various webpages is constructed in advance; a data acquisition mode matched with the webpage is obtained from a webpage data acquisition knowledge base, and the data acquisition mode defines that an analysis mode of each field in the webpage is analyzed based on multi-modal features (at least two of DOM structure fingerprint features, visual position features and semantic features of the fields); according to the method, the analysis accuracy of the webpage data can be greatly improved, so that stable and efficient execution of a webpage data acquisition task is guaranteed.

The invention relates to the technical field of visual position recognition, and particularly discloses a multi-modal visual position recognition reordering method and system based on guidance, and the method comprises the steps: obtaining a query image, and retrieving a plurality of candidate images based on a pre-trained visual basic model and the query image; constructing a composite multi-modal prompt object, wherein the composite multi-modal prompt object comprises an image pair formed by the query image and the current candidate image, and an instruction text used for guiding a multi-modal large language model to perform visual comparison; outputting a structured similarity judgment result, wherein the result comprises a quantitative similarity score; and sorting based on the similarity scores corresponding to all the candidate images, and determining the candidate image with the highest score as an optimal matching result. Through combination of guiding type prompt engineering and structured output, an intermediate text generation link is avoided fundamentally, and the calculation efficiency is improved while the fidelity of all original visual information is reserved.

Systems and methods of the present disclosure are directed to a method for training a machine-learned visual attention model. The method can include obtaining image data that depicts a head of a person and an additional entity. The method can include processing the image data with an encoder portion of the visual attention model to obtain latent head and entity encodings. The method can include processing the latent encodings with the visual attention model to obtain a visual attention value and processing the latent encodings with a machine-learned visual location model to obtain a visual location estimation. The method can include training the models by evaluating a loss function that evaluates differences between the visual location estimation and a pseudo visual location label derived from the image data and between the visual attention value and a ground truth visual attention label.

The invention relates to the technical field of security and protection monitoring, and discloses a security and protection monitoring alarm linkage processing method and system. The method comprises the following steps: constructing a three-dimensional scene graph of a monitoring area; obtaining visual position information and wireless positioning information of a target, and executing constrained Kalman filtering fusion to obtain a target position; mapping the target position to the three-dimensional scene graph for threat assessment to obtain an intelligent alarm signal; and performing linkage control in the three-dimensional scene graph based on the intelligent alarm signal to obtain an equipment activation instruction and a blocking strategy. According to the method, high-precision three-dimensional target positioning is realized in a complex environment, an alarm decision strategy can be dynamically optimized according to historical statistical data, the problem of high false alarm rate caused by a fixed threshold mode is effectively reduced, and the stability and the reliability in different environment conditions are improved.

The invention discloses a big data-based universe scenic area multi-modal interaction special effect generation method and system, and relates to the technical field of virtual reality, and the method comprises the steps: obtaining universe scenic area information, obtaining historical tourist information according to the universe scenic area information, and generating a universe scenic area multi-modal interaction special effect according to the historical tourist information; and obtaining historical interaction time information and historical interaction tourist information corresponding to each interaction special effect. According to the method, the preferred tourists of each interaction special effect are accurately analyzed through the preferred tourist feature information, a data basis is provided for subsequent interaction special effect generation, the interaction deviation angle model is constructed through the special effect visual position and the tourist position, the special effect interaction degree is calculated in combination with the interaction time, single interaction duration judgment is replaced, and the interaction efficiency is improved. The tourist preference analysis accuracy is improved, the appropriate interaction special effect is selected by screening the interaction special effects corresponding to the meta universe scenic spots, the special effect individuation adaptation degree is improved, and the tourist immersion experience is optimized.

The invention provides an in-vehicle target positioning method and device, a vehicle and a storage medium, the method is applied to the field of sound source localization, and the method comprises the steps: carrying out the hand image collection and audio collection of a user in the vehicle, and the audio collection is executed by a plurality of audio collectors; determining an acoustic position coordinate of the hand of the user according to a first acquisition time when each audio acquisition device acquires the sound; according to the first acquisition time, the acquired hand image and the acquisition time of the hand image, determining a visual position coordinate of the hand of the user and a confidence coefficient of the hand of the user in executing the preset gesture; and performing filtering correction on the acoustic position coordinates according to the visual position coordinates and the confidence coefficient. The method can effectively reduce the positioning error caused by a single sensor, is high in positioning precision, is high in anti-interference capability, and is suitable for various complex in-vehicle environments.

The invention relates to the technical field of full-automatic cup protruding, in particular to an intelligent full-automatic cup protruding system which comprises a shell, a sample plate feeding mechanism, a code scanning mechanism, a clamping table mechanism and a discharging box are sequentially installed above the shell from left to right, and a visual recognition mechanism is distributed above a visual placement station. Through cooperation of the sample plate feeding mechanism, the code scanning mechanism, the suction plate transferring mechanism, the opening mechanism, the clamping table mechanism, the cup protruding station device and the automatic cup protruding mechanism, the suction plate mechanism controls a sample plate to rotate by 90 degrees to be placed on a cup protruding station, and a sample plate pressing head is driven by a gear to press the sample plate. When the gear reaches a fixed torsion, the torsion limiter in the gear slips, so that a sample plate is pressed for cupping, the functions of testing other items such as film thickness, gloss, adhesive force and the like are added to part of advanced models, and operation and results can be remotely controlled, quickly exported, optimized and perfected through networking.

The embodiment of the invention relates to an indoor positioning and tracking method and system based on UWB and visual fusion and a storage medium, and the method comprises the steps: firstly carrying out the real-time detection through employing a camera, obtaining a plurality of visual position points, determining a target candidate point in the plurality of visual position points based on a position track, and carrying out the positioning and tracking of the target candidate point; determining a visual moving direction based on the visual track of the target pedestrian and the target candidate point, determining whether the visual moving direction is matched with the track direction of a first UWB track, if yes, updating the position point of the position track at the current moment based on the target candidate point, and if not, updating the position point of the position track at the current moment based on the target candidate point; and if yes, updating the position point of the position track at the current moment based on the position point of the position track at the previous moment and the position point of the first UWB track at the current moment. According to the method, visual positioning and UWB positioning are fused, track direction matching is carried out, a position track is updated based on a matching result, and the accuracy of positioning and tracking is improved.

The present invention discloses a sample-based selective knowledge distillation method and system. First, a network is trained to segment SEG images and RGB images to complete VPR tasks respectively. Then, high-quality knowledge contained in SEG features is transferred to RGB features. While the invariant features in the segmented images are strengthened in the RGB features, there is no need to retain additional branch networks during testing. The segmented images are input into the training network through weighted unique hot encoding to extract scene structure information for visual position recognition tasks. Through sample-based selective knowledge distillation, different distillation loss weights are applied according to importance. Through knowledge distillation, the invariant features in the additional modal information are strengthened in the RGB features, and high-quality knowledge is strengthened in a directionally manner, thereby greatly improving the network accuracy. Finally, when testing the network, only RGB images are used for global retrieval, which takes into account the real-time, high-precision and environmental robustness of the system.

To provide a technique capable of specifying a visual position on a surface of an inspection target without using a camera capable of eye tracking.SOLUTION: A visual inspection support device for supporting a visual inspection of an inspection target by an inspector comprises: a light source which irradiates light toward the inspection target from near the eyes of the inspector; an imaging unit which is disposed near an illumination unit that illuminates the inspection target and images reflected light that has been irradiated from the light source and reflected by the inspection target; a specification unit which specifies a trajectory of a reflection position of the reflected light on the surface of the inspection target by using an image captured by the imaging unit; and a determination unit which determines, with reference to the specified trajectory, an inspected region and an uninspected region of the surface.SELECTED DRAWING: Figure 1

The application provides a visual position recognition method based on attention compression coding features. A hierarchical database is established based on a scene three-dimensional map; and global features are extracted by using an attention mechanism and codebook compression coding based on the hierarchical database to realize online position recognition. The application is used to reduce the calculation overhead of feature extraction in offline and online stages, improve global recall accuracy, and thus guarantee the service quality of user positioning.

The application belongs to the technical field of coil packaging, and particularly relates to a micro-fan coil sealing structure, a coil packaging process and a micro-fan. The micro-fan coil sealing structure comprises a sealing cover body, a plurality of independent sealing cavities for accommodating coils are uniformly and interval arranged along the circumferential direction on one side of the sealing cover body close to the coils; a positioning structure for fixing the coils is arranged in the sealing cavities. The independent sealing cavities are arranged on the sealing structure of the coils to provide independent mounting space and positioning reference. Meanwhile, the positioning structure is arranged in the sealing cavities, so that the radial / circumferential position of the coils in the sealing cavities is more easily limited. When batch assembly is performed, the randomness of'manual visual placement' can be reduced, and the consistency of the coil mounting position is improved. The stable position of the coils means that the relative position of the coils and the rotor magnetic steel / magnetic circuit is more stable, so that the problems of uneven magnetic field, torque fluctuation and the like caused by eccentricity and inclination are reduced. Furthermore, the fan operation is stable, and the probability of abnormal vibration or efficiency fluctuation is reduced.

The invention provides a virtual reality video interaction method and device, electronic equipment and a storage medium. According to the invention, a video display sphere associated with the position of a head-mounted display device is established in a virtual reality scene, and the video display sphere is used for playing a panoramic video containing a dynamic object and maintaining a fixed picture reference orientation; determining the real-time pose of the dynamic object under the current video playing progress from the pose change data sequence of the dynamic object during video production; setting a dynamic interaction body in a coordinate system corresponding to the video display sphere according to the real-time pose, and maintaining a synchronization relationship between the dynamic interaction body and the video display sphere; and when detecting that an interaction event occurs between the user interaction operation and the dynamic interaction body, generating interaction feedback in an area corresponding to the visual position of the dynamic object in the panoramic video picture.

The invention relates to a data set enhancement method for visual position identification, which comprises the following steps of: firstly, processing a Nordland data set containing complete seasonal image distribution of all places by utilizing a Conditional GAN, and realizing image seasonal conversion based on a Pix2pix model; then, the obtained 12 Pix2pix models are used for supplementing season information missing in some places for an existing GSV-Cimes data set; images in the obtained GSV-Ciences data set with more complete season information are input into a feature extractor for training; and carrying out performance analysis on the obtained trained image feature extraction model, and comparing with the current several VPR algorithms. Results show that the method provided by the invention has more excellent performance when seasons change and illumination and weather change.

The invention provides a visual position identification method of a pre-training model based on feature aggregation and selection, and belongs to the technical field of visual position identification. According to the method, global feature descriptors and local feature descriptors are extracted, an adaptive feature aggregation technology is adopted, and the extraction of landmark information with discrimination is emphasized, so that the difference between a pre-training model and an actual visual position identification application is effectively made up. And for the global feature descriptors, integrating image global information into descriptors with rich semantics through a feature aggregation method by adopting a DINOv2 + ViT model structure. For extraction of local feature descriptors, a patch-level descriptor extraction module is designed, and dense and distinctive local features can be output. The features play an important role in two-stage tasks, firstly, the features are used for local matching of candidate images, and then the matching precision is further improved through a reordering strategy.

The application discloses a cross-domain cross-source data alignment method and system and electronic equipment. The method first inputs a plurality of sets of table data to be aligned. Then, key-value pairs in the data and positions of the key-value pairs in the table are extracted. Next, a data multi-modal representation model is used to generate vector expressions of the keys, values and visual positions. Semantic distances of the vector expressions from different data are calculated. Finally, the semantic distances between different data are evaluated to determine an alignment result. In addition to using pairing between keys, the application considers matching of values, enhancing matching of keys in the prior art. In addition to the representation in the text, the application fuses a table visual structure as part of semantic representation of the key-value pairs, breaking the limitation of the prior art that only uses single-modal information for matching.

The invention provides a video fusion method, system and device and a program product, and the method comprises the steps: obtaining a real-time video stream to be fused to a three-dimensional scene model, and pre-calibrating a display plane and a three-dimensional scene reference point in the three-dimensional scene model; determining a coordinate of a visual position point of a three-dimensional scene reference point on a display plane under a visual angle of a current world camera corresponding to a current frame original image in the real-time video stream; mapping the visual position point into a calibrated pixel coordinate range to obtain a coordinate of a target image reference point; calculating a coordinate transformation relation between the original image and the target image according to the coordinate of the target image reference point and the calibrated coordinate of the original image reference point; performing transformation processing on the current frame original image based on the corresponding coordinate transformation relation to obtain a corresponding target image; and rendering the target image on the display plane. According to the invention, a video fusion angle adaptive mechanism is added, and the video fusion effect is effectively improved.

The invention relates to the technical field of industrial Internet of Things and edge intelligence, and provides an intelligent tool tracing method and system. According to the implementation scheme, the method comprises the steps of sensing an instrument occurrence event and constructing each observation data node; generating a spatial connectivity hypothesis based on visual similarity among the observation data nodes so as to dynamically construct a tool movement trajectory diagram; clustering the environmental visual features of all observation data nodes to construct an environmental cognitive network; in response to triggering of a tool search event, confirming a search starting point based on a multi-starting-point competition mechanism; performing probability diffusion calculation based on multi-source behavior evidence fusion on the environment cognitive network by taking the search starting point as an initial source, and generating probability distribution of the target tool in each position in the network; based on the probability distribution, a search instruction is generated, and the search instruction comprises visual position information and semantic information of the target tool. According to the embodiment of the invention, rapid positioning and intelligent traceability of the tool can be realized under the condition of no preset map.

The application relates to the field of artificial intelligence and discloses a position recognition method and device based on entropy-controlled hierarchical attention, equipment and a medium, which comprises the following steps: receiving an input image to be recognized and performing overlapping block embedding processing to generate an initial overlapping token sequence; performing hierarchical multi-scale feature extraction on the initial overlapping token sequence, dynamically regulating the attention distribution of each layer by information entropy during feature extraction to obtain preliminary features of each layer; performing feature refining and integrating processing on the preliminary features of all layers to output corresponding scene descriptors; performing scene matching in all preset scenes according to the scene descriptors, and outputting the position information of the input image according to the scene matching result. The application can be applied to business scenes such as financial technology and medical health, dynamically regulates the attention distribution of each layer based on information entropy during hierarchical multi-scale feature extraction, adaptively focuses on the most discriminative scene area for position recognition, and improves the accuracy of visual position recognition.

The invention relates to a method and system for out-of-domain detection. Methods, systems, and aircraft for performing image analysis to assist in refueling operations are disclosed herein. The method includes receiving a 2D image from a camera, determining a domain score for the 2D image based on previously defined training data, and transmitting the 2D image to a visual position estimation system in response to the domain score being greater than a predetermined threshold, thereby creating a transmitted 2D image.

A unified place recognition framework handles both retrieval and re-ranking with a unified transformer model. The re-ranking modules utilizes feature correlation, attention value, and x / y coordinates into account, and learns to determine whether an image pair is from a same location.

This invention provides a wafer arrangement method and system for improving the wavelength uniformity of MOCVD epitaxial wafers. The wafer arrangement method for improving the wavelength uniformity of MOCVD epitaxial wafers includes: measuring and calibrating the wavelength offset of each groove on the carrier; calculating the predicted wavelength offset of the wafer to be processed using a pre-built wavelength test model; sorting the wafers to be processed in the current batch according to their predicted wavelength offsets; sorting the grooves on the carrier in the opposite manner to their predicted wavelength offsets; then matching the sorted wafers with the sorted grooves in sequence; mapping the matching results back to the original physical number of the groove to generate a visual placement guide, which is then used by an operator or robot to perform the corresponding placement under the guidance of the visual placement guide.

The invention discloses a register-assisted aggregation visual position identification method based on DINOv2, and the method comprises the following steps: 1, collecting streetscape images in a data set, forming a training sample and a test sample, and preprocessing the training sample and the test sample; step 2, constructing a neural network comprising a multi-scale feature fusion module according to a DINOv2 network; putting the preprocessed training sample into an initial model, obtaining a multi-scale feature through the neural network, and obtaining a local descriptor at the same time; 3, training the neural network, and optimizing parameters of the neural network to obtain a trained model; and 4, inputting a test image in the test sample into the trained model to obtain a verification image matched with the test sample, and outputting a recall rate. According to the method, the representation capability of features under different resolutions is enhanced, the robustness of global descriptors is improved, and the accuracy of position retrieval is improved.

The invention discloses a positioning method and system for an automatic overlapping machine of honeycomb core aramid paper, the positioning system comprises a visual detection system, a visual position moving system, a position compensation mechanism and a control system, and the positioning system is characterized in that the visual detection system is arranged on the visual position moving system; the control system includes a recipe management system and can drive the visual position movement system and the position compensation mechanism. According to the method, image information of overlapped honeycomb core aramid paper is collected through the visual detection system, the image information of the overlapped honeycomb core aramid paper and a pre-stored reference pattern are compared through the control system, the position of the overlapped honeycomb core aramid paper is automatically adjusted, and high-precision overlapping is achieved. Position parameters are pre-stored in a formula management system of the control system, automatic position adjustment of the visual detection system is achieved when honeycomb core aramid paper of different specifications is replaced, the debugging time is saved, and the debugging material loss is reduced. Through the improvement, the production efficiency can be improved on a large scale, and the overlapping production of multi-specification and large-scale honeycomb core aramid paper is realized.

According to particular embodiments, one process includes retrieving a multi-view interactive digital media representation that includes numerous images fused together into content and context models. The process next includes retrieving and processing audio data to be integrated into the multi-view interactive digital media representation. A first segment of audio data may be associated with a first position in the multi-view interactive digital media representation. In other examples, a first segment of audio data may be associated with a visual position or the location of a camera in the multi-view interactive digital media representation. The audio data may be played in coordination with the multi-view interactive digital media representation based on a user's navigation through the multi-view interactive digital media representation, where the first segment is played when the first position or first visual position is reached.

The application provides a kind of facility agriculture picking robot vision-haptics fusion control method, the method comprises: by obtaining the RGB-D image and laser radar point cloud of facility environment, extract crop ridge center line, real-time solution chassis linear velocity vector and angular velocity vector;Obtain the three-dimensional point cloud and surface image of fruit, solve the maturity grade of fruit, and construct visual stiffness prior model, estimate the elastic modulus of fruit;According to the elastic modulus on-line mapping generation mechanical arm end impedance controller's target stiffness matrix and damping matrix;Adopt the position-based visual servo strategy, the pose error between end effector and target fruit is converted into the expected angular velocity of each joint of mechanical arm, drive end to approach target;When the contact force detected by end six-dimensional force sensor reaches the preset threshold, the system is switched from pure visual position control to adaptive impedance control mode, and the visual planning path is dynamically corrected using force feedback.

The invention relates to the technical field of machining centers, in particular to a vertical turning, milling and grinding integrated combined machining center which comprises a base, two first electric sliding rails, a portal frame, a second electric sliding rail, a controller, a combined machining unit, an eccentric rotating workbench, a position precise control mechanism, a pose adjusting mechanism and a visual position detector. Wherein the two first electric sliding rails are fixedly installed on the left side and the right side of the top of the base respectively, the portal frame is fixedly installed at the output ends of the two first electric sliding rails, the second electric sliding rail is fixedly installed on the top of the portal frame, and the controller is fixedly installed on the front side of the base. Parallel operation of machining and pre-positioning is achieved, the defect of insufficient precision caused by coordinated control of multiple servo systems is overcome, and precision accumulative errors caused by multiple times of clamping are eliminated; the method has the outstanding advantages of being high in machining efficiency, stable in positioning precision and high in adaptability.

Embodiments of the present application disclose a visual-based in-vehicle sound source positioning method, a controller and a computer readable storage medium. The method comprises: acquiring visual data, determining a visual position of a target person based on the visual data, determining a plurality of visual noise penalty parameters based on the visual data, generating a visual covariance matrix based on the plurality of visual noise penalty parameters, performing Kalman filtering processing on the visual position and the visual covariance matrix based on a preset Kalman filter, and obtaining a sound source position of the target person. In the embodiments of the present application, the plurality of visual noise penalty parameters are determined based on the visual data, so as to measure the interference of various noise factors on the sound source position from multiple dimensions, the visual covariance matrix is generated by comprehensively considering the plurality of visual noise penalty parameters, and finally, the Kalman filtering processing is performed on the visual position and the visual covariance matrix by using the Kalman filter, so as to obtain a sound source position with high reliability, thereby improving the positioning accuracy and robustness.