477 results about "Real image" patented technology

The invention relates to the technical field of deep counterfeit image detection, and discloses a deep counterfeit detection method and system, a storage medium and computer equipment. The method comprises the following steps: firstly, constructing a reference data set containing a forged image and an original real image; secondly, through an integrated model, generating antagonistic samples for the reference data set, and integrating the successfully attacked antagonistic samples into an antagonistic sample set; and finally, merging the reference data set and the adversarial sample set, and constructing a robustness enhanced data set containing four types of samples. In the model training stage, multi-classification cross entropy loss and comparative learning loss are combined, and expression of the model in a feature space is optimized through comparative learning constraint, so that the model learns discriminative features with more compact intra-class features and more dispersed inter-class features. The model trained by the method not only can effectively defend against attack and improve robustness, but also surpasses original detection performance on clean samples, and has remarkable technical advantages and application value.

The invention provides an optical lens, which comprises four lenses with focal power and sequentially comprises a first lens with negative focal power, a second lens with negative focal power, a third lens with positive focal power, a fourth lens with negative focal power, a fifth lens with negative focal power and a sixth lens with negative focal power from an object side to an imaging surface along an optical axis, the object side surface of the second lens is a convex surface, and the image side surface of the second lens is a concave surface; the object side surface of the third lens is a convex surface, and the image side surface of the third lens is a convex surface; the image side surface of the fourth lens is a convex surface; wherein the maximum field angle FOV of the optical lens and the real image height IH corresponding to the maximum field angle of the optical lens meet the following conditions: 35 degrees / mmlt; fOV / IHlt; 46 degrees / mm; the focal length f3 of the third lens and the focal length f4 of the fourth lens satisfy 0.5 lt; f3 / f4lt; and 0.8. According to the optical lens provided by the invention, through specific surface shape matching and reasonable focal power distribution, the lens has one or more advantages of an ultra-wide angle, a large aperture, high imaging quality, high collimation and the like.

The invention discloses a visual detection optimization control method, device and equipment based on digital twinning and a storage medium, and relates to the technical field of visual detection, and the method comprises the steps: obtaining an analog image and a real image in a digital twinning environment, and carrying out the frequency domain feature transformation of the analog image and the real image, thereby obtaining an image spectrum feature; a frequency domain alignment model is established based on multi-band spectrum envelope guide residual mapping, and the structure of virtual and real image spectrum features is kept aligned; further extracting features through multi-scale convolution and channel dependence mapping to obtain virtual-real fusion features; quantifying channel similarity and establishing a covariance regularization constraint, performing channel correction on the cross-domain features, and eliminating feature drift to obtain second virtual-real fusion features; and finally, performing visual detection and micro defect identification based on the features. The problem that a virtual sample and a real sample are different in local texture structure and channel distribution is solved.

The invention provides a camera calibration method and device, electronic equipment and a storage medium, and belongs to the technical field of data processing, and the method comprises the steps: obtaining laser radar point cloud data, inertial measurement unit data and reference camera image data of a vehicle, and obtaining initial calibration parameters of a to-be-calibrated camera; based on the laser radar point cloud data, the inertial measurement unit data and the reference camera image data, constructing a point cloud map containing color textures and a pose track of a vehicle, and converting the point cloud map into a surface grid model; according to the pose track and the initial calibration parameters, the surface grid model is rendered to a visual angle of the to-be-calibrated camera, and a rendered image and a rendered depth map corresponding to the to-be-calibrated camera are generated; based on the real image data, the rendered image and the rendered depth map, establishing a corresponding relationship between two-dimensional image features and three-dimensional space points in the real image data; and constructing a re-projection error objective function based on the corresponding relationship, and optimizing the external reference and the internal reference of the camera to be calibrated by minimizing the objective function.

The invention discloses a three-dimensional reconstruction method, device and equipment based on 3D Gaussian. The method comprises the steps of performing data preprocessing based on a three-dimensional point cloud output by sparse reconstruction of a target scene, a camera internal reference and a camera pose; initializing and training a 3D Gaussian model based on a real image, the three-dimensional point cloud after candidate object mask screening, a camera pose, a depth information graph and a scale alignment parameter to obtain a coarse-grained 3D Gaussian model of the target scene; generating a dynamic target mask by taking a rendered image generated by rendering the target scene based on the coarse-grained 3D Gaussian model as a reference and combining the real image and the candidate object mask; and on the basis of the coarse-grained 3D Gaussian model, the dynamic target mask and all input data obtained by initializing and training the coarse-grained 3D Gaussian model, carrying out refined reconstruction and optimization on the coarse-grained 3D Gaussian model. According to the method, unification of high precision, high efficiency and high robustness is realized, and the increasing actual industrial application requirements are met.

The invention discloses a dynamic scene reconstruction method, computer equipment and a program product. According to the method, picture information and camera poses of the same scene are obtained, a standard field composed of three-dimensional Gaussian primitives is constructed in a three-dimensional space, a deformation field is constructed, the standard field and the deformation field at different times are subjected to joint rendering in a three-dimensional Gaussian splashing mode, and various parameters are trained and optimized according to the difference between a synthetic image and a real image. Performing motion perception according to the motion condition of the three-dimensional Gaussian primitives, taking the primitives with smaller motion as static primitives, and performing partition modeling on the primitives with larger motion in time to obtain a layered dynamic scene model; and calling the model for rendering according to the given time and the camera pose in the reasoning stage, and outputting a target image. According to the method, the reconstruction precision of the high-dynamic target can be improved while the rendering efficiency is ensured, and the continuity and visual quality of the dynamic scene in the time dimension are improved by introducing the time consistency constraint at the boundary of the time subinterval.

According to the time series prediction method based on the multi-modal contrast learning technology, original multivariable time series data are converted into structured visual representation and language representation, multi-modal representation with consistent inner performance can be constructed without depending on external natural language or real image data, and the time series prediction method is high in practicability. The deep semantic understanding capability of the model on the complex operation state of the rail transit is effectively enhanced; a multi-modal contrast learning mechanism is introduced, visual and text modal representation is aligned in a shared embedding space, positive sample consistency is maximized through InfoNCE loss, negative sample interference is suppressed, and the robustness and generalization ability of time sequence features are remarkably improved; and the importance of each variable on a prediction task is dynamically evaluated by using the aligned multi-modal representation, and key variables are automatically screened, so that the redundant information interference is reduced, and the prediction precision and the calculation efficiency of the model in a high-dimensional multi-variable scene are also improved.

In a biometric identification system, a transparent body can have a curved platen surface. The curved platen surface can be removably contacted by an anatomy of a user. The anatomy can have an anatomical pattern. An illumination source can illuminate the curved platen surface with an illumination beam. The illumination beam can propagate away from the curved platen surface as a return beam having a beam pattern shaped to correspond to the anatomical pattern. An imaging lens can focus the return beam to form a real image of the beam pattern at an image location. Field-curvature-correcting sensor optics can include a sensor. The sensor can sense the real image of the beam pattern at the image location. The field-curvature sensor optics can include a planar sensor and an image-flattener lens. The field-curvature sensor optics can include a curved, convex sensor.

The invention discloses a physical mechanism embedded industrial data generation method and device and a storage medium, and belongs to the field of artificial intelligence and industrial production. The method comprises the following steps: firstly, acquiring process parameters of a target object, inputting an improved condition diffusion model, and outputting a corresponding generated microscopic structure image; the improved condition diffusion model comprises a double-flow condition encoder and a strain adaptive physical loss function; the double-flow condition encoder is used for encoding the thermal deformation temperature and the dependent variable and fusing the thermal deformation temperature and the dependent variable into a condition vector; the strain adaptive physical loss function is used for constraining the consistency of the generated image and the real image in the frequency domain amplitude spectrum and edge gradient features in the model training process; inputting the generated microscopic structure image into a recrystallization characteristic evaluation module based on an auxiliary segmentation network to obtain a recrystallization characteristic quantitative evaluation result of the generated microscopic structure image; and the auxiliary segmentation network is used for carrying out recrystallization region segmentation on the microscopic structure image and calculating a recrystallization fraction based on a segmentation result.

The invention discloses an orbit foreign matter detection method and system, and relates to the technical field of orbit detection.The method comprises the steps that an unknown foreign matter model is built through Gaussian distribution and a convex hull algorithm, and a known foreign matter model is built in combination with prior data; extracting a flat area of the ballast bed, projecting the foreign body model into a depth map and generating a shadow, and superposing to obtain a synthetic image with a mark; training a segmentation network by using the synthetic image and a real image, and obtaining a potential foreign matter mask through thermodynamic diagram binaryzation, connected domain analysis and size screening during reasoning; and inherent facility false alarms are eliminated through a secondary classification network, a real foreign matter detection result is finally output, alarm is triggered according to a detected threshold value, and efficient and accurate detection is realized. According to the method, on the premise that real foreign matter samples are scarce and ballast bed scenes are complex, efficient, accurate and real-time ballastless track foreign matter detection is achieved, meanwhile, the false alarm rate is reduced, and the generalization ability of the model to unknown foreign matter is improved.

Systems, apparatus, articles of manufacture, and methods to generate synthetic images for training of defect detection models are disclosed. An example system disclosed herein produces synthetic images of pallet defects to train object detection models. In some examples, a small set of real images containing defects, associated masks and textual descriptions is used to fine tune a latent diffusion model. The fine-tuned model accepts a masked input image, a mask that defines the region to be altered, and a defect description, and generates a synthetic image with the defect inpainted into the masked region. In some examples, generated synthetic images are filtered to remove outliers that do not match the real defect distribution. The filtered synthetic dataset, together with a limited set of real images, is then used to train a downstream object detection model capable of identifying pallet damage in captured images.

The invention provides an optical lens, which comprises seven lenses with focal power and sequentially comprises a first lens with negative focal power, a second lens with positive focal power, a third lens with negative focal power, a fourth lens with negative focal power, a fifth lens with negative focal power and a sixth lens with negative focal power from an object side to an imaging surface along an optical axis, the object side surface of the second lens is a concave surface, and the image side surface of the second lens is a convex surface; the third lens has negative focal power; the object side surface of the fourth lens is a convex surface, and the image side surface of the fourth lens is a convex surface; the fifth lens has positive focal power, and the image side surface of the fifth lens is a convex surface; the object side surface of the sixth lens is a concave surface, and the image side surface of the sixth lens is a concave surface; the seventh lens has negative focal power; the optical total length TTL of the optical lens, the real image height IH corresponding to the maximum field angle of the optical lens and the maximum field angle FOV of the optical lens meet the following conditions: 0.17 lt; 1 degree * TTL / (IH / 2) / (FOV / 2) lt; and 0.24. According to the optical lens provided by the invention, the imaging quality of the optical lens can be improved, the aberration is reduced, and the imaging quality of the optical lens is improved.

The invention provides a scene modeling method and device for electric power facilities, equipment and a medium. The method comprises the following steps: acquiring a plurality of frames of real images shot on the electric power facility at a plurality of visual angles; acquiring point cloud data of the electric power facility by adopting an SFM technology, and obtaining a plurality of frames of rendering images by adopting a 3DGS algorithm; for each frame of rendered image, calculating the gradient of each Gaussian body in the rendered image through back propagation based on the difference between the rendered image and the real image under the corresponding view angle; when the gradient value of any Gaussian body is greater than a preset gradient threshold value, adjusting the Gaussian density of a space region corresponding to the Gaussian body according to a covariance matrix after projection of the Gaussian body and a texture structure type at a position corresponding to the Gaussian body in a real image corresponding to the rendered image; and obtaining a scene model of the electric power facility based on the adjusted Gaussian density. The method is used for achieving the effect of improving the modeling precision of the electric power facility.

The application relates to a metal product automatic welding equipment control method and system. For a metal product to be welded, a target image is collected, a defect detection model is used to detect defects of the metal product to be welded according to the target image, and then welding control of the automatic welding equipment is performed according to the defect information. Since the control of the automatic welding equipment considers the defect problem of the metal product, the problem of poor welding effect of the metal product with defects can be solved. Moreover, since the pre-trained defect detection model is trained according to real images and simulation images, the model has strong generalization ability and can be applied to defect detection in various scenes, improving the accuracy of defect detection. Therefore, the technical scheme controls the automatic welding equipment to weld the metal product through defect information with high accuracy, improving the welding effect of the metal product.

The invention discloses a two-dimensional irregular accessory identification method based on depth image retrieval, and belongs to computer vision. The method comprises the following steps: firstly, analyzing the inner and outer contours of an accessory from a CAD drawing, rendering a standard drawing on a canvas with a fixed resolution ratio, and generating a multi-view rendering graph according to a hemisphere discrete pose as simulation data to replace a camera image so as to form a standard-simulation sample pair; then, ResNet18 combined with double attention is adopted to form a twin network, contrastive learning is matched, the same accessory is aggregated in a feature space, different accessories are separated, and vectors are extracted from a standard image in an off-line mode to construct an index; and finally, coding a real image into a query vector in an online stage, performing neighbor retrieval by combining inner product or cosine similarity in a vector library, and returning a Top-K candidate. According to the scheme, both precision and efficiency are considered, and good robustness is achieved for view angle transformation and detail deformation. Under the scene of scarcity of real data or high labeling cost, the training and retrieval performance can be maintained by relying on pre-construction of multi-view simulation data and a vector library.

The invention relates to the technical field of computer vision, in particular to a remote sensing image panchromatic sharpening method and system based on a double-domain diffusion model, and the method comprises the steps: obtaining randomly generated Gaussian noise, inputting the Gaussian noise into a pre-training double-domain diffusion model, and obtaining a reconstruction difference image of a real image and a multispectral image in a reverse sampling stage; adding the reconstructed difference image and the multispectral image to generate a high-resolution multispectral image; compared with a traditional method, the method has remarkable advantages in the aspects of noise suppression and detail enhancement, and a multispectral image with more sufficient spectral information and spatial information can be generated.

The invention discloses a dynamic nerve radiation field real-time three-dimensional reconstruction method and device based on window attention and gradient equilibrium, and the method comprises the steps: carrying out the stratified sampling of rays, and obtaining sampling points; projecting the sampling points to six groups of two-dimensional feature planes; querying corresponding feature vectors from the six groups of feature planes through bilinear interpolation, and generating spatio-temporal features through Hadamard product fusion; inputting the fused spatial-temporal characteristics into a decoding network of a Swin Transform, and outputting the color and the volume density of a corresponding sampling point; and carrying out volume rendering integration along the ray of the camera to generate a composite image, and optimizing and updating the model according to a multi-target loss function between the composite image and a real image. The invention aims to solve the problems of large memory consumption, poor rendering quality, long training time and poor geometric consistency in the prior art, and realizes real-time three-dimensional reconstruction of a dynamic scene by using six-plane grid representation, Swin Transform feature generation and a gradient equilibrium mechanism.

A machine learning method according to this embodiment trains a recognizer using a composite image, acquires a labeled real image based on an image captured by a sensor, stores, when at least one of the results of the recognition when the labeled real image is input to the recognizer match the label, results of the recognition performed by the recognizer, performs machine learning using a data set group including a plurality of the data sets, a real image and a composite image forming a pair in each of the data sets, thereby generating an image transformer, generating a labeled real image as a result of the image transformer transforming the composite image into the real image, and trains the recognizer based on the labeled real image.

A system for training a generative machine learning model, comprising: a hardware processor, configured for: extracting a plurality of real primitive features from a plurality of real images captured by a sensor in a physical environment; extracting a plurality of other primitive features from a plurality of other images depicting another physical environment; training a generative machine learning model to produce a refined image in response to a synthetic image, the training using the plurality of real images, the plurality of real primitive features, the plurality of other images and the plurality of other primitive features, where the generative machine learning model is adapted for receiving the plurality of real primitive features and additionally or alternatively the plurality of other primitive features as input, to produce a trained model; and providing the trained model to another hardware processor for the purpose of generating synthetic training data.

This invention discloses an imaging device, comprising an imaging module and a mounting base. The imaging module includes a housing, a flat lens, a display, and a first slip ring. The flat lens has a first side and a second side, with the display located on the first side. Light emitted from the display passes through the flat lens on the second side to form a floating real image. The first slip ring is located at the bottom of the housing, and the display is connected to an external power source via the first slip ring. The imaging module is rotatably mounted on the mounting base. A second slip ring is provided on the surface of the mounting base facing the imaging module. When the imaging module is mounted on the mounting base, the second slip ring contacts the first slip ring to achieve electrical connection. The second slip ring is adapted to be electrically connected to an external power source. Thus, the imaging module is rotatable relative to the mounting base, allowing it to be used independently or in conjunction with the mounting base. When in conjunction with the mounting base, the imaging module can rotate 360° relative to the mounting base and can stop at any angle, facilitating user observation of the floating real image.

The invention discloses a target detection method and system based on a CAD model, and belongs to the technical field of image recognition, a three-dimensional CAD model of a to-be-detected target is taken as a detection basis, dependence on a real image and annotation data as a training sample is completely worsened, so that the training sample is not needed, and when a new object needs to be detected, the detection efficiency is greatly improved. Target detection can be carried out only by providing a CAD model, expensive data acquisition and labeling processes are not needed, the expandability and application flexibility of the technology are greatly improved, a classification verification mechanism is set in a final result, and therefore the reliability of a detection result is guaranteed, and the detection efficiency is improved. The problems that the detection cost is too high and the application expansibility is poor due to the fact that a large number of real images and manual annotations need to be collected to train a detection model before the target is detected in an existing target detection method are solved.

The application discloses a kind of based on identity recognition probability distribution's fake face swapping image detection method and device, construct real image and corresponding identity label, identity label is carried out smoothing operation to construct smooth identity label;Real image is used to carry out multiple rounds of identity label-based supervision learning to identity recognition model to optimize parameter;Copy current parameter optimization identity recognition model as copy model, based on the pixel area of greatest contribution in the prediction result of copy model in real image is shielded to construct shielded image;Shielded image is used to carry out multiple rounds of smooth identity label-based supervision learning to current parameter optimization identity recognition model to optimize parameter;Until training ends, the maximum prediction probability value output by identity recognition model is compared with threshold value, when maximum prediction probability value is less than threshold value, it is detected as fake face swapping image.This method prevents attack while improving the detection accuracy and generalization of fake face swapping image.

The present invention addresses the problem of providing an image display device and AR glasses that are both capable of displaying a virtual image so as to be superimposed on a real image, are compact, and have excellent luminance uniformity of the virtual image. The image display device according to the present invention includes at least: a display element; a diffraction element including an optically anisotropic layer for changing the angular characteristics of luminous intensity in an image displayed by the display element; a beam splitter that reflects at least a portion of the image in which the angular characteristics of luminous intensity were changed by the diffraction element; and a half mirror that reflects, toward the beam splitter, at least a portion of the image reflected by the beam splitter. The optically anisotropic layer has a liquid crystal alignment pattern including a region in which the orientation of an optical axis derived from a liquid crystal compound changes while continuously rotating along at least one in-plane direction.

The invention provides an optical lens, which comprises six lenses in total, and sequentially comprises a first lens with negative focal power, a second lens with negative focal power, a third lens with negative focal power, a fourth lens with negative focal power, a fifth lens with negative focal power and a sixth lens with negative focal power from an object side to an imaging surface along an optical axis, the object side surface of the second lens is a convex surface, and the image side surface of the second lens is a concave surface; the third lens has positive focal power, and the image side surface of the third lens is a convex surface; the object side surface of the fourth lens is a convex surface, and the image side surface of the fourth lens is a convex surface; the fifth lens has negative focal power, and the image side surface of the fifth lens is a concave surface; the object side surface of the sixth lens is a convex surface, and the image side surface of the sixth lens is a convex surface; wherein the real image height IH corresponding to the maximum field angle of the optical lens and the effective focal length f of the optical lens meet the following conditions: 2.1 lt; iH / flt; and 2.6. According to the optical lens provided by the invention, through specific surface shape matching and reasonable focal power distribution, the lens has one or more advantages of ultra-wide angle, high pixel, large aperture, small distortion, high imaging quality and the like.

The small-size lens defect classification detection system comprises a light source imaging system, a defect simulation data set and a defect detection and identification system, the light source imaging system comprises a first light source located under the lens, a second light source surrounding the horizontal periphery of the lens, a third light source projecting from the side of the lens at a low angle and having an irradiation angle lower than that of the second light source, and an image acquisition device for acquiring front or cylindrical surface images of the lens; generating lens defect images in batches by the defect simulation data set; the defect detection and identification system takes a lens defect image and a real image collected by the light source imaging system as training data, and after model training, defect identification and classification are carried out on the lens image collected by the light source imaging system in real time. The defect imaging contrast ratio is effectively improved through multi-light-source combined illumination, the problem that defect samples are scarce is solved in combination with simulation data, and efficient, accurate and automatic detection of multiple defects of a small-size lens is achieved.

The invention discloses a three-dimensional image display device and a vehicle, and is applied to the technical field of vehicle-mounted display. The three-dimensional image display device comprises a first image generation module and a second image generation module, the first image generation module is used for generating and displaying a first image, the second image generation module is used for generating and displaying a second image, and the second image generation module comprises a second image generation unit with transmittance; the positions of the first image and the second image in the first direction do not coincide, and a three-dimensional visual image which can be presented through the second image generation unit is formed. According to the invention, the human eyes can watch the second image and the first image which are not overlapped through the second image generation unit along the first direction, and the second image and the first image are real images which are not overlapped and distorted, so that a real three-dimensional visual image can be presented when the human eyes watch the image, and no vertigo feeling is generated.

The invention provides an optical lens, which comprises four lenses with focal power and sequentially comprises a first lens with positive focal power, a second lens with negative focal power, a third lens with positive focal power, a fourth lens with negative focal power, a fifth lens with positive focal power and a sixth lens with negative focal power from an object side to an imaging surface along an optical axis, the second lens has positive focal power, the object side surface of the second lens near the optical axis is a convex surface, and the image side surface of the second lens is a convex surface; the third lens has positive focal power, the object side surface of the third lens is a concave surface near the optical axis, and the image side surface of the third lens is a convex surface near the optical axis; the fourth lens has negative focal power, the object side surface of the fourth lens is a concave surface near the optical axis, and the image side surface of the fourth lens is a convex surface near the optical axis; wherein the real image height IH corresponding to the maximum field angle of the optical lens and the effective focal length f of the optical lens meet the following conditions: 1.3 lt; iH / flt; 1.6); the curvature radius R7 of the object side surface of the fourth lens and the curvature radius R8 of the image side surface of the fourth lens meet the following formula:-0.45 lt; (R7-R8) / (R7 + R8) lt; and-0.2%. According to the optical lens provided by the invention, the overall aberration of the optical lens can be reasonably corrected, and the imaging quality of the optical lens is improved.

This invention relates to an array optical system, an apparatus equipped with an array optical system, and an assembly and adjustment method, belonging to the field of optical imaging, and solves the problems of small field of view and large front-end diameter in existing optical systems. An array optical system includes an imaging primary mirror and an image-rotating mirror group; the primary mirror is used to transmit visible light rays to form a primary imaging visible light real image plane, and the image-rotating mirror group is used to transmit visible light rays to form a secondary imaging visible light real image plane. This invention solves the problems of small field of view and large front-end diameter in existing optical systems.

The invention discloses a self-supervised image denoising method and system combining local and global features, and belongs to the field of image processing. The method comprises the following steps: constructing a Patch-level mask mechanism convolution module PLMC, and processing an input image by adopting PLMCs with convolution kernels of different sizes; constructing a local feature extraction branch used for carrying out pixel shuffling down-sampling operation on a feature map output by the PLMC with the convolution kernel of the first size and then carrying out local feature extraction; constructing a global feature extraction branch used for carrying out pixel shuffling downsampling operation on a feature map output by the PLMC with the convolution kernel of the second size and then carrying out global feature extraction; and carrying out element addition on the local features and the global features, then carrying out pixel shuffling up-sampling operation, and recombining into a de-noised image with the same size as the original image as a final reconstructed de-noised image. The denoising effect of the real image is improved.

This invention discloses a method and system for detecting coating defects on the surface of rollers based on image generation, belonging to the field of industrial product quality inspection technology. The method includes the following steps: acquiring continuous image data of the roller during a stable rotational speed phase and a sudden increase in rotational speed, recording the start time of the speed change, and generating rotational speed change identification information; organizing the continuous images in chronological order based on the rotational speed change identification information, comparing the brightness distribution of the images before and after the speed change, extracting areas of sudden brightness increase, and generating a set of bright areas. This invention achieves precise correspondence between image acquisition and roller operating state by introducing rotational speed change identification information, eliminating interference from sudden changes in illumination on texture recognition; and through dynamic backtracking of the texture real image set and peeling risk information, it achieves continuous tracking and adaptive detection of coating defects, improving recognition stability and detection reliability.