314 results about "Image View" patented technology

The invention discloses a large-scene three-dimensional reconstruction method based on three-dimensional Gaussian sputtering, and relates to computer graphics. The method comprises the following steps: collecting a multi-view image set of a large scene; obtaining a scene sparse point cloud according to the multi-view image set; performing monocular depth estimation on the multi-view image by using a pre-trained depth prediction network to obtain monocular depth estimation priori; the method comprises the following steps of: performing global training on a scene by utilizing scene sparse point cloud and monocular depth estimation prior to obtain an initial three-dimensional Gaussian model, and performing space grid division on the initial three-dimensional Gaussian model to obtain a plurality of scene blocks with axis alignment bounding boxes; setting image view angle data of each scene block; performing deep supervised training on the Gaussian ellipsoids in the plurality of scene blocks by using a parallel GPU (Graphics Processing Unit); combining the trained scene blocks to obtain a final three-dimensional Gaussian model; in view of low geometric structure reconstruction precision caused by only depending on color information of a multi-view image in large-scene three-dimensional rendering, the method improves the reconstruction precision of large-scene rendering.

The invention discloses a radio frequency interference positioning method and device combining a brightness temperature image and a subspace, and the method comprises the steps: obtaining an original scene brightness temperature image for observing an earth scene, constructing a simulation background brightness temperature image of the earth scene, and obtaining a brightness temperature image after the first SNR is improved; carrying out image processing on the brightness temperature image of the original scene, generating a radio frequency interference (RFI) mask corresponding to the brightness temperature image of the original scene, and obtaining a brightness temperature image after the SNR is improved for the second time; and constructing a covariance matrix of the brightness temperature image after the second SNR improvement, carrying out subspace decomposition on the covariance matrix, and carrying out geolocation on an RFI source observed in an observation image view field. According to the method provided by the invention, the positioning precision of the RFI source under the condition of low signal-to-noise ratio is improved.

The invention discloses a power transmission line channel hidden danger distance measurement method based on point cloud image fusion. The method comprises the steps of data joint calibration, target recognition and modeling of a power transmission line channel scene, target segmentation and positioning of the power transmission line channel scene and judgment of the distance between a hidden danger target and a wire. The method has the advantages that the point cloud data and the image data are fused, the space coordinate conversion relation between the point cloud data and the image data is calibrated and solved, the point cloud data are projected to the image view angle to generate a new point cloud projection mapping graph, training recognition is conducted through the image data and the point cloud data, the recognition accuracy of the hidden danger target is improved, the SAM model serves as a basic model, and the recognition efficiency is improved. The point cloud projection mapping graph and the hidden danger target in the image data are used as samples, training is carried out, a hidden danger target segmentation model is generated, accurate contour information of the hidden danger target is extracted, and effective early warning information is provided for the safety of a power transmission line channel.

The invention provides a self-adaptive perception positioning method based on unsupervised fusion BEV, and the method comprises the following steps: collecting point cloud data of environment information nearby a vehicle in real time through employing a laser radar, and collecting image view data of the environment information nearby the vehicle in real time through employing a camera; the method comprises the steps of converting multi-view image view data and point cloud data into the same BEV space, respectively generating camera BEV features and point cloud BEV features, fusing the camera BEV features and the point cloud BEV features in a projection mode to obtain fused BEV features, performing semantic segmentation and instance segmentation on the fused BEV features, and generating a vectorized fused BEV map; learning query is introduced through query-based collaborative self-adaption, interaction between multi-view image view data and the fused BEV map and collaborative self-adaption of the multi-view image view data and the fused BEV map are achieved, and a spatial perception result is generated; and self-checking the spatial perception result, checking the rationality of the image, and completing the closed loop of the perception positioning process. According to the invention, the fusion BEV under unsupervised learning can rapidly detect and obtain the environment information, and autonomous verification is carried out, so that the accuracy of vehicle positioning is ensured, and the experience feeling of the vehicle is improved at the same time.

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.

A method includes generating, using at least one processing device, a ground truth optical flow map for displacement of pixels within a reference image based on a motion model. The motion model determines 3D coordinates of pixels within the reference image based on estimated depths of the pixels within the reference image. The method also includes performing, using the at least one processing device, 3D to 2D reprojection of the pixels within the reference image based on the motion model to generate a reprojected image view corresponding to a shifted camera perspective. The method further includes generating, using the at least one processing device, an occlusion mask for the reference image. The occlusion mask corresponds to occluded pixels within the reprojected image view. In addition, the method includes performing, using the at least one processing device, occlusion region inpainting of the occluded pixels to generate an inpainted reprojected image view.

Reproduction 1.6: reference image view showing only the part intended for design registration; reproductions 2.3 and 3.3: enlarged image views of the part intended for design registration; reproductions 2.4 and 3.4: enlarged image views showing the changed state of the part intended for design registration; the parts shown in the broken lines do not form part of the claimed designs; designs nos. 2 and 3: the dash-dot lines represent the boundary between the claimed design and unclaimed design and form no part of the claimed design; reproductions 1.2, 2.2, and 3.2 are the image views showing second image in a sequence; reproduction 1.3 is the image view showing third image in a sequence; reproduction 1.4 is the image view showing fourth image in a sequence; reproduction 1.5 is the image view showing fifth image in a sequence; reproduction 1.6 is the reference view showing only the part intended for design registration; reproductions 1.6, 1.7, 2.5, 2.6, 3.5 and 3.6 are for reference purposes only and form no part of the claimed designs.

A medical apparatus (100) includes a medical device (108) with bending sections and a distal end, at least one imaging device (110) at the distal end, at least one sensor (112), and at least one processor (201) which performs receiving input data input data from user input, the imaging device (110), the sensor (112), or combinations thereof, determining a bending plane and a bending angle of the distal end of the medical device (108), predicting location movement and position of the bending sections based on input data, displaying an image view based on the input data and displaying the predicted location movement and position of the bending sections on the image view.

An imaging device for capturing images through an eyepiece of an optical viewing device. The imaging device includes a camera that aligns with the eyepiece for capturing the images viewed through the eyepiece, and a display for displaying the images captured by the camera. The imaging device determines a type of the optical viewing device. The imaging device presents a workflow on the display based on the type of the optical viewing device. The imaging device captures metrics related to user interactions with the workflow presented on the display and the images captured by the camera during the workflow.

An information processing method is an information processing method executed by one processor or executed by a plurality of processors in cooperation, the information processing method including a first acquisition step for acquiring map information, a second acquisition step for acquiring current position information of a flight vehicle, a third acquisition step for acquiring information concerning a virtual viewpoint for a user to check the flight vehicle in an image, and a generation step for generating a virtual viewpoint image, which is an image viewed from the virtual viewpoint, based on the map information, the current position information of the flight vehicle, and the information concerning the virtual viewpoint.

The invention relates to the technical field of composite materials, in particular to a long glass fiber reinforced plastic preparation process autonomous regulation and control method based on image vision, which comprises the following steps: judging the color difference interval of glass fiber and a standard sample, and judging whether floating fiber exists in colloidal particles or not according to the color difference interval; judging whether the floating fiber condition is in a first floating fiber degree or a second floating fiber degree according to the floating fiber ratio; in the first fiber floating degree, judging that the fiber floating reason is insufficient and non-uniform impregnation according to the diameter difference between the material diameters; judging whether the equipment parameters are in a preset range, and adjusting the equipment parameters; in the second floating fiber degree, the deviation condition is judged according to the diameter difference value of the materials; and judging whether the offset degree is a first category or a second category according to the offset, analyzing the offset trend, and adjusting the proportion or traction force of the polypropylene resin. By performing image view analysis on the colloidal particles prepared from the long glass fiber reinforced plastic, the reason of fiber floating is judged, and the preparation process is correspondingly regulated and controlled.

The invention relates to the field of intelligent glasses, and provides an image information acquisition method and device, intelligent glasses and a storage medium, and the method comprises the steps: obtaining an image view finding instruction; determining view-finding range information according to the image view-finding instruction, wherein the view-finding range information is used for indicating a view-finding area and a non-view-finding area in a view field range of the intelligent glasses; adjusting the light transmittance of the intelligent glasses lens based on the view-finding range information, so that the first light transmittance of the intelligent glasses lens and a first area corresponding to the view-finding area is greater than the second light transmittance of the intelligent glasses lens and a second area corresponding to the non-view-finding area; and obtaining target image information corresponding to the view finding area based on the image information obtaining instruction. The intelligent glasses display the view-finding area and the non-view-finding area in the view field range based on different light transmittances, so that a user can simply and visually distinguish the view-finding range of the camera of the intelligent glasses, the user can clearly know the picture captured by the camera, and the user experience of the intelligent glasses for image shooting is improved.

The invention relates to the technical field of image processing, in particular to a panoramic image splicing method for a curtain wall building facade. The method comprises the following steps: performing three-dimensional scanning on a curtain wall building facade to obtain entity point cloud data of the facade, and selecting a plurality of discrete points from the entity point cloud data as reference anchor points; shooting a plurality of local images of the curtain wall building facade, and in the shooting process of each local image, recording the spatial correspondence between the image viewing range and the reference anchor point; and according to the spatial correspondence, aligning the image area corresponding to the reference anchor point in each local image to the position of the reference anchor point in the entity point cloud data to form a preliminary spliced image. According to the invention, through an image processing technology, zero-offset binding of the panorama and the BIM is completed through one-time shooting, so that the millimeter-level positioning precision and the inspection efficiency of the curtain wall facade are improved.

The invention provides an anti-interference target detection method and system based on automatic driving scene multi-modal fusion, and the method comprises the steps: employing a semantic segmentation and instance segmentation combination method based on a camera image in an image view coding layer, and employing a Mask-RCNN model, and precisely judging a possible shielded object and region; in a feature processing link, a time self-attention mechanism is introduced, a feature map is weighted from a time dimension, an occluded object is focused, irrelevant information is inhibited, and the capability of capturing features of the occluded object is enhanced; besides, a caching mechanism is arranged in the model, a target confidence coefficient change method is adopted, the detection confidence coefficient of continuous video frames is stored, and the shielding condition is recognized by analyzing the confidence coefficient change condition of a target object in the continuous video frames. According to the invention, the detection precision and reliability of the shielded object are improved, and a more efficient and more accurate solution is provided for the object detection of the unmanned driving technology.

The invention discloses a single-classification industrial control system anomaly detection method based on double-view information bottleneck fusion, which belongs to the technical field of industrial control system anomaly detection and comprises the following steps of: in a training stage, simultaneously constructing a time sequence view and a pseudo image view; respectively extracting characterization through a time dynamic encoder and a pre-trained visual backbone; performing cross-view compression in the variational information bottleneck fusion module and establishing projection branches only depending on sequence representation; a normal subspace is learned around a center vector, an accurate hypersphere is established to describe normal data, and a center distance is used as an abnormal score and a quantile threshold is used to complete discrimination; in the detection stage, online recognition can be performed only by inputting a time sequence view. According to the method, dual-view prior and variational information bottleneck are fused, end-to-end collaborative optimization is carried out on the dual-view prior and the variational information bottleneck and deep support vector data description, an accurate decision boundary can be established under the condition of a small number of samples, and a solution for industrial control system anomaly detection under the condition that the samples are limited is provided.

3D image data of a skeletal portion is acquired. While a portion of a tool is disposed at a first location along an insertion path, first and second x-rays are acquired from first and second image views. A processor registers the x-rays with the 3D image data, identifies a location of a tool within the x-rays, and computes an anticipated forward path of the tool within the 3D image data from the x-rays. The tool is moved to a second location along the insertion path and an additional x-ray is acquired from a single image view. The processor facilitates identifying if the tool has deviated from the anticipated forward path by registering the additional x-ray to the 3D image data and identifying a location of the portion of the tool within the additional x-ray. Other embodiments are also described.

The invention provides a new goods category discovery method based on asymmetric enhancement and confidence collaborative learning, relates to the technical field of computer vision and artificial intelligence, and aims to improve the recognition capability of the new goods category discovery method on known and unknown goods in a dynamic retail environment. According to the method, image views with different intensities are generated through an asymmetric data enhancement strategy, and the adaptability of a model to the appearance change of goods is enhanced. Meanwhile, a confidence-guided collaborative learning mechanism is adopted, high-confidence samples in unmarked data are dynamically utilized, and noise interference is reduced. In addition, aiming at the long tail distribution problem of a known category, a Logit adjustment strategy is introduced, and the classification effect is optimized. By integrating various loss functions, a multi-task learning framework is constructed, accurate recognition of known goods and automatic discovery and clustering of unknown new goods are realized, and the intelligent level and practicability of the intelligent electronic scale are remarkably improved.

The invention discloses a multi-modal depth fusion 3D target detection method based on reliable depth estimation, and the method comprises the steps: firstly constructing an image feature extraction network and a millimeter-wave radar point cloud feature extraction network, and respectively extracting image features and millimeter-wave radar point cloud features; secondly, fusing the image and millimeter wave radar point cloud features to obtain depth probability distribution, multiplying the depth probability distribution by the image features to obtain image view cone features, making a depth probability distribution label by using the laser radar point cloud after depth completion, and performing supervision training based on the depth probability distribution; then combining the view cone features of the image and the point cloud features of the millimeter wave radar, and obtaining a feature FA by using a Splat network and a deformable attention mechanism; and finally, based on the feature FA, completing 3D target detection by using a 3D target detection head, and outputting a detection result. According to the method, depth estimation is optimized through the millimeter wave radar, so that the accuracy and robustness of 3D target detection are improved.

The invention relates to the technical field of lightweight rendering, in particular to a cloud-edge coordinated point cloud model lightweight rendering processing method and system. The method comprises the following steps: performing point cloud coordination prior processing according to electric power facility point cloud data to generate coordination prior point cloud data; a point cloud rendering model is constructed, point cloud curved surface coefficient calculation and curved surface reconstruction processing are carried out, and a point cloud curved surface reconstruction rendering model is generated; obtaining image view angle data, performing rendering model pose alignment processing and model texture enhancement mapping processing, and generating an enhanced texture rendering model; performing rendering coordination lightweight deployment and lightweight real-time rendering processing on the enhanced texture rendering model to generate lightweight real-time rendering data; and obtaining user behavior data, performing point cloud data rendering feedback analysis and cloud edge coordination caching in combination with the lightweight real-time rendering data, and feeding back to the terminal. According to the method, rendering of point cloud power equipment data curved surface reconstruction can be lightened, and redundant transmission of cloud edge rendering data is reduced and coordinated.

The invention discloses a multi-modal rumor detection system based on multi-field perception, relates to the technical field of rumor detection, and is used for solving the technical problem of low multi-field rumor detection precision in the prior art. The rumor detection system comprises a multi-view feature extraction module, a multi-domain perception module, a multi-level knowledge fusion module and a final classifier. Wherein the multi-view feature extraction module is used for respectively extracting corresponding single-mode features from a text mode and an image mode, and acquiring aligned image and text features as multi-mode view features; the multi-domain sensing module is used for learning cross-domain shared knowledge and domain specific knowledge at a text view angle, an image view angle and a multi-mode view angle; the multi-level knowledge fusion module is used for fusing cross-domain shared knowledge and domain specific knowledge of a text view angle, an image view angle and a multi-mode view angle; and the final classifier is used for obtaining a prediction result according to the cross-domain shared knowledge and the domain specific knowledge obtained by the multi-level knowledge fusion module.

A remote operator terminal is used by a remote operator for a remote operation of a moving body. A first image is captured at a first timing by a camera mounted on the moving body from a first point of view. A second point of view is one at a second timing later than the first timing. A homography process converts the first image viewed from the first point of view into a second image viewed from the second point of view based on perspective projection transformation using a projection plane associated with the camera. The remote operator terminal acquires a combination of a plurality of second images by applying the homography process using a common projection plane to a combination of a plurality of first images, and displays the combination of the plurality of second images on a display.

The invention relates to the technical field of video collection, in particular to a video synchronous collection method and device based on an optical far image, and provides the following scheme: obtaining a calibration image of to-be-synchronized content and a user posture parameter; calculating a geometric mapping matrix and a point spread function through the calibration image and a preset calibration sequence, and performing geometric predistortion and frequency domain compensation on the to-be-synchronized content to obtain a compensation frame; and combining with the rotation range predicted by the attitude parameters, generating a sub-frame sequence by adopting multi-view rendering of time multiplexing, and outputting the sub-frame sequence by a display end, so as to realize stable far-image watching experience under the condition of small desktop size. According to the scheme, visual fatigue caused by close gaze can be relieved, and the intelligent level of learning interaction is improved.

To make it easier for safety managers to check the site conditions. [Solution] The safety support system 100 includes an object detection unit 13 that detects moving images of objects that meet detection conditions from among multiple video sources MI10 of each detection area 200 captured by multiple imaging devices (cameras 110) registered in a database 30; an alarm condition setting unit 12 that registers detection conditions; a detection moving image extraction unit 17 that extracts and saves moving images of objects that meet detection conditions as detection moving images MI11 from among the video sources detected by the detection conditions; and an extracted moving image viewing unit 18 that displays multiple detection moving images of the same detection area on a display device 3a in synchronization with the time of imaging.

The present invention relates to the field of image processing and medical technology, and specifically to a full-slice pathological image segmentation and auxiliary annotation method. First, the full-slice pathological image WSI is divided into a model training data set and a data set to be annotated, the cancer area in the model training data set is annotated, and a pathological image segmentation data set and a model prediction data set are constructed; the pathological image segmentation data set is used to train the U‑Net segmentation model; secondly, the trained U‑Net segmentation model is used to predict the small image blocks in the model prediction data set, and the segmentation results of each small image block are generated and pieced together into the result of the whole pathological slice; finally, the contour of the segmentation result is extracted, and the coordinate point values ​​of the contour are written into an XML annotation file, and the pathologist can review the annotation quality on the pathological image viewing software ASAP. The full-slice pathological image segmentation and auxiliary annotation method of the present invention can effectively segment the cancer area, and the auxiliary annotation can further reduce the annotation burden of the pathologist.

An optical viewing system includes an optical viewing device having an eyepiece. The system further includes an imaging device that includes a housing having a bracket that is configured to removably attach the imaging device to the optical viewing device. The imaging device includes a camera configured to capture images viewed through the eyepiece and a display screen configured to display the images captured by the camera. The system further includes a secondary display screen physically separate from the optical viewing device and the imaging device. The secondary display screen pairs with the imaging device to display the images viewed through the eyepiece of the optical viewing device. The pairing between the imaging device and the secondary display screen expires when a condition is satisfied.

Most of the internet traffic today is carried out via the Transmission Control Protocol (TCP). The main advantage of TCP is that it provides reliable data transfer to the application layer and simplifies programming. The protocol maximizes data throughput but may also lead to noticeable transmission delay in wide area networks (WAN). A client-server based medical image viewing system is disclosed that achieves high data throughput over TCP without impacting responsiveness. Special timestamp messages inserted into the data stream allow the system to detect situations where network latency increases noticeably and to obtain a reliable estimate of sustained transfer bandwidth. The system applies a feedback scheme that avoids network delays by limiting send bandwidth. In addition other parameters, in particular image compression settings, are dynamically adjusted depending on current network quality.

The invention relates to the technical field of liquid crystal display, and provides liquid crystal display equipment based on a field sequence and a color image display method. White backlight is inserted in a liquid crystal response stage to reduce a color separation phenomenon, and the white backlight is inserted in the liquid crystal response stage, so that white light can be emitted at the time when the original backlight is turned off; according to the backlight module, the original backlight opening time in the color backlight stage is not occupied, so that the backlight source of the backlight module is always in the lightened state, the backlight light efficiency is improved, meanwhile, the backlight source of the backlight module can be prevented from being converted between the two light and dark states of closing and lightening, the image flickering problem is solved, and the image watching quality is improved.

A computer-implemented method is provided for facilitating the acquisition of ultrasound imaging data over an extended period of time. with improved consistency in the image view which is captured. A first set of ultrasound imaging data, comprising one or more images representing one or more views of the anatomy of a patient, is acquired and stored in a datastore. This image data is then used as reference data for subsequent image acquisition, wherein subsequent image data is compared in real time with the reference image data to obtain a measure of a degree of correspondence between the views depicted. This can be used to guide the acquisition of the new image data so that it is consistent with the previous image data. For example. when the view represented in the newly acquired image data sufficiently matches that in the reference data. an alert message can be issued to the user.

Provided by the present disclosure are a display device and a display method therefor. By detecting the position of a viewer, the display angle of the scene in an image to be displayed is adjusted according to the position of the viewer so as to adapt to the viewing angle of the viewer. Therefore, the image viewed by the viewer can better conform to the real objective scene, and when the viewing angle changes, different angle pictures of an object can be viewed, thereby improving the display effect.

The present disclosure relates to an ophthalmic lens for a wearable head-up display, the ophthalmic lens comprising: an optical center, a primary gaze axis, and a holographic optical element having an apex wherein the holographic optical element is integrated with the ophthalmic lens such that the apex coincides with an image viewing axis to allow viewing of digital content, and wherein the optical center of the lens is positioned relative to the primary gaze axis and the apex is positioned relative to the optical center. Also disclosed is a wearable display comprising an eyeglass lens mounted in the frame, an eyeglass frame, and an image source mounted on an arm of the frame adjacent to the eyeglass lens to project light onto a holographic optical element of the eyeglass lens to generate one or more image eyespots.