38 results about "Edge detector" patented technology

The invention provides an evidence obtaining method and system based on image processing, and the method comprises the following steps: S1, generating a pixel-level depth-of-field distribution diagram of an input image through a multi-scale encoder-decoder network, employing an edge perception optimization layer in a decoding stage, and improving the depth-of-field boundary precision through minimizing a local gradient consistency loss function; s2, performing depth-of-field rationality verification based on an optical imaging physical model, and triggering a first-level tampering alarm by calculating a defocusing fuzzy radius and a gradient direction of a selected region when a difference between the defocusing gradient directions of a target region and a background region exceeds a preset threshold value; and S3, dynamically positioning a key pixel region, identifying a depth-of-field mutation boundary by using an edge detector, calculating by combining local texture complexity, screening a pixel set of which the entropy value is higher than a threshold value and which is located at the mutation boundary, correlating metadata to verify the rationality of the physical size and the spatial position of an object, and eliminating false detection caused by perspective transformation.

A photodetection device according to the present disclosure includes a plurality of light-receiving sections that each generates a pulse signal including a pulse corresponding to a result of light reception, a plurality of edge detectors that each generates a detection signal by detecting an edge of the pulse in the pulse signal generated by a corresponding light-receiving section, and an adder that generates a detection value indicating number of the pulses on the basis of a plurality of the detection signals. The edge detectors each include a first latch circuit that generates a first signal by latching the pulse signal on the basis of a first clock signal, a second latch circuit that generates a second signal by latching the first signal on the basis of a second clock signal that is an inverted signal of the first clock signal, a combination circuit that generates a third signal on the basis of the pulse signal, the first signal, and the second signal, and a third latch circuit that generates the detection signal by latching the third signal on the basis of the first clock signal.

A computer-implemented image processing method is provided. The computer-implemented image processing method includes: obtaining a pair of training samples, the pair including a training image with a first sharpness and a reference image with a second sharpness greater than the first sharpness, wherein at least a portion of the training image and at least a portion of the reference image in the same pair have the same content; inputting the training image into an image enhancement convolutional neural network to generate a training enhancement image; inputting the training enhancement image into an edge detector; generating a plurality of first edge maps by the edge detector; inputting the reference image into the edge detector; generating a plurality of second edge maps by the edge detector; and adjusting parameters in the image enhancement convolutional neural network to minimize at least one or more of the first and second losses.

The invention provides a non-contact air suspension ultrathin wafer turning device, which comprises a machine body, two clamping jaws, an edge detector, a displacement sensor and a control system, and is characterized in that the two clamping jaws can move relative to the machine body, each clamping jaw is provided with a crack for placing a wafer, and the edge detector is arranged in the crack; a plurality of air holes for spraying air towards the wafer are formed in the crack, so that an air film is formed on the outer side of the wafer to enable the wafer to suspend in the crack, the edge detector is arranged in the crack and used for monitoring the real-time position of the edge of the wafer, the displacement sensor is used for detecting the position state of the wafer, and the control system controls the clamping jaw to rotate when the clamping jaw turns over. And calculating execution air pressure based on data of the edge detector and the displacement sensor so as to adjust air film pressure distribution, so that the wafer is kept at a preset position. According to the non-contact air suspension ultrathin wafer turning device provided by the embodiment of the invention, in the wafer turning process, the pollution risk and physical damage of the wafer can be reduced.

The invention belongs to the technical field of building construction, relates to a ceramic tile seam beautifying method and an automatic ceramic tile seam beautifying robot, and aims to solve the problems that manual seam beautifying depends on experience, the efficiency is low, the cost is high, and the quality is unstable. The method comprises the following steps: positioning and moving equipment (a center detector is aligned with a tile gap, and an edge detector detects an obstacle avoidance area in real time); automatically pasting a film (releasing the film by a film storage roller, attaching the film by a film pasting roller, pressing the film by a film pressing roller and exposing the gap by a film cutting ring cutter, and cutting the film by a film cutter); and automatically gluing and scraping glue (lifting to drive a glue gun to inject glue. The robot comprises a moving trolley, a follow-up film pasting mechanism and a gluing mechanism, the moving trolley comprises a detection device and a multi-wheel walking device, the follow-up film pasting mechanism and the gluing mechanism are arranged, full-automatic seam beautifying can be achieved, manpower is reduced, and efficiency and quality are improved.

The application discloses an image internal occlusion relationship prediction method and system based on edge information guidance. The method comprises two parts of an edge detector and an occlusion relationship predictor, and comprises the following steps: inputting an image through an edge detector composed of an image segmentation module, a plurality of pooling Transformer modules and a deconvolution module to generate an edge detection graph; then in the occlusion relationship predictor, the edge detection graph acts on the image element features generated by the edge detector through an edge information guidance module; then the new image element features are sequentially optimized through a plurality of pooling Transformer modules; and finally, a last deconvolution module adjusts and enlarges the finally obtained image element features in resolution, and outputs an occlusion relationship graph. The application first proposes a new method of guiding feature calculation by the generated edge detection graph and directly generating an occlusion relationship graph. Compared with other existing methods, the application has less calculation amount and better prediction effect.

The invention discloses a multi-information progressive fusion image colorization method based on a deep convolutional neural network. The method is composed of five core modules: an input module is responsible for receiving an original image and carrying out preprocessing operations such as graying and normalization; the feature extraction module is actually a VGG network and is responsible for extracting multi-level features of the image; the multi-information fusion module adopts a channel attention mechanism to collaboratively fuse gray features, edge features extracted by a Canny edge detector and semantic features extracted by a DeepLabv3 + semantic segmentation network; the progressive colorization generation and optimization module comprises a plurality of U-Net sub-networks, and realizes colorization image generation from coarse to fine through a staged generation strategy in combination with joint optimization of L1 loss, perception loss and adversarial loss; and the output module carries out reverse normalization processing on the result. According to the method, the problems of color distortion and detail loss in the prior art are effectively solved, the color image with accurate color, rich details and vivid vision can still be generated in a complex scene, and the method can be widely applied to the fields of old film color restoration, medical image enhancement and the like.

A document reading apparatus includes a possible document edge detector that detects, on the basis of a density change indicated by image data acquired by a reading device, a density-changing point, where image density is equal to or higher than a threshold, and most downstream in a transport direction, at each of positions in a main scanning direction, as a possible position of a document edge, and detects an image formed of pixels continuously aligned in the main scanning direction at the possible position, as a possible document edge, and a document size detector that detects a document size in the main scanning direction, on a basis of the image of the possible document edge, when the image of the possible document edge detected by the possible document edge detector is formed of a single image extending in the main scanning direction at a same position in the transport direction.

A system for cutting blanks is provided. The blanks include a first blank and a second blank. The system comprises an edge detector configured to determine a configuration of an edge of the first blank; a cutter configured to cut an edge of the second blank; and one or more processors operatively connected with the edge detector and the cutter, the one or more processors controlling the cutter to cut the edge of the second blank into a configuration based on the detected configuration of the edge of the first blank.

Automatic target evaluation system for bullet holes, comprising a camera unit, a computing unit and image analysis software, wherein the camera unit continuously records images of a target and the computing unit is designed to extract contours and edges using a Canny edge detector, to perform a circle approximation using area boundaries and to identify possible bullet holes.

The application provides a train wheel position monitoring method, device and system, which can accurately measure the position of the train wheel by periodically using a stereo camera and a thermal imager, and automatically display the measurement image, thereby significantly improving the accuracy and robustness of the measurement. When an image selection instruction is received, the system automatically marks the image to be measured and the train wheel to be measured in the image, then processes the image through an edge detector to extract the vertical edge of the train wheel, and further uses an iterative algorithm to accurately extract the horizontal coordinate of the edge white point. The arc angle between the horizontal center of the image and the edge white point is calculated to determine the width of the train wheel, and the offset is compared with a preset template to determine whether the train wheel is displaced. This method not only reduces the dependence on camera settings, improves the operation efficiency, and reduces the implementation and operation cost, but also meets the strict requirements of the railway transportation system for high safety and high efficiency, and has great popularization value.

The application discloses a method capable of storing different magnifications and automatically setting boundary strength, adopts a projection system, and comprises a workbench, an objective lens, a projection screen, a frontmost mirror and a last mirror; the objective lens comprises a mounting seat and a lens; a plurality of sensing elements are arranged on the mounting seat, the sensing elements are connected with an edge trigger controller, a trigger element is arranged on the lens, and the trigger element automatically triggers the corresponding sensing element after the lens is mounted on the mounting seat, so that the edge trigger controller automatically distinguishes the magnification of the lens, a plurality of different magnifications are preset and stored, and corresponding boundary strengths are respectively set; during operation, the corresponding boundary strengths are automatically adjusted according to the lens with different magnifications, manual setting is not needed, the setting magnification is avoided to be forgotten, and the edge detector is used in combination to automatically enable the edge detector to perform edge detection by using different threshold values according to different magnifications, so that the precision and convenience are considered.

The application relates to the technical field of object recognition, and discloses an image recognition detection method and system based on weak supervision learning, which comprises the following steps: S1, generating initial pseudo labels of a to-be-detected image through point annotation, an edge detector and a self-adaptive flood filling algorithm; S2, constructing a point supervision salient object detection model based on a visual converter; S3, performing the first round of training on the point supervision salient object detection model to obtain a final salient map; and S4, suppressing non-salient objects of the final salient map, performing the second round of training on the point supervision salient object detection model, and completing the recognition of the image. The application solves the problem that the existing image recognition detection technology can only cover part of an image, cannot ignore invalid objects, leads to low efficiency, and has the characteristics of strong supervision capability.

The application discloses a dynamic edge detection device and a preparation method and a dynamic edge detection device thereof. The dynamic edge detection device comprises oppositely arranged first and second substrates and a ferroelectric liquid crystal layer. The first substrate is provided with a first transparent electrode layer on one side, and the second substrate is provided with a second transparent electrode layer on one side. The first substrate is provided with an orientation layer on the side facing the second substrate. The ferroelectric liquid crystal layer comprises ferroelectric liquid crystal molecules forming a layered helical structure, and the optical axis presents a circularly changing arrangement and an angular distribution. In the case of applying an external electric field perpendicular to the first substrate and with a voltage greater than or equal to a threshold voltage, the helical structure of the ferroelectric liquid crystal layer is disentangled. The dynamic edge detection device provided by the application has electrically controllable adjustable optical axes. When linearly polarized light is incident, the outgoing vector polarization rotation can be realized by changing the polarity of the applied electric field, thereby realizing the edge switching in the orthogonal dimension, and high-speed and high-efficiency one-dimensional optical image edge detection in any direction is realized.

The application relates to the technical field of CT imaging equipment, and discloses a CT system and a security inspection equipment. The CT system comprises a rectangular channel extending along the Z direction, a rotating assembly, a ray source and a detector assembly. The detector assembly comprises a middle detector and an edge detector, the rays of the ray source pass through the rectangular channel and irradiate the detector assembly, when the rotating frame rotates by 360 degrees, an annular edge detection area for detecting the rectangular channel is formed between the ray source and the edge detector, the overlapping part of the annular edge detection area and the rectangular channel is not greater than the non-overlapping part of the annular edge detection area and the rectangular channel; the number of rows of crystal modules on the first sub-edge detector and the second sub-edge detector is half of the number of rows of crystal modules on the middle detector; the crystal modules of the first sub-edge detector and the second sub-edge detector are staggered along the Z direction and have no overlapping position.

The utility model relates to the technical field of optical lens detection, and discloses an optical lens corner detection tool, which comprises a Y-axis linear module, a vacuum chuck, a rotary motor and an optical lens edge detector, the Y-axis linear module is used for feeding optical lenses, and a Y-axis moving sliding table of the Y-axis linear module is provided with a containing disc matched with the optical lenses to be placed. The vacuum chuck is used for adsorbing and fixing the optical lens on the placing disc, the rotary motor is used for driving the vacuum chuck and the adsorbed and fixed optical lens to rotate, the rotary motor is mounted on the XZ-axis linear module, and the XZ-axis linear module acts to drive the vacuum chuck and the rotary motor to move to a detection station; and the optical lens edge detector is used for carrying out corner detection on the optical lens on the detection station. Automatic feeding, fixing and rotating of the optical lens can be achieved, it is ensured that a detection window is accurately aligned with a detection station, manual intervention is reduced, and labor cost is reduced.

The invention discloses a fracture X-ray film AI auxiliary diagnosis system. The system adopts a deep learning architecture based on a convolutional neural network, and comprises an image preprocessing module, a deep learning recognition module, a diagnosis result output module and a user interaction interface module. The image preprocessing module carries out de-noising processing by combining adaptive median filtering with a bilateral filtering technology, and an improved Canny edge detector is combined with a deep learning edge enhancement network to identify skeleton features. The deep learning recognition module constructs a feature extraction network based on a ResNet-50 pre-training model, and adopts a weighted cross entropy loss function and an Adam optimizer for training. And the diagnosis result output module generates a visual fracture line label by applying a gradient weighted class activation mapping technology, and automatically generates a structured report containing fracture types, severity evaluation and diagnosis and treatment suggestions. The system supports DICOM and JPG format image input, the diagnosis accuracy reaches 78%-80%, and the fracture diagnosis efficiency and accuracy are remarkably improved.

The application relates to a 3DGS-based transparent object reconstruction and new view synthesis method, which comprises the following steps: S1. reconstructing a dense three-dimensional point cloud of a data set and camera pose information according to an image sequence; S2. using a three-dimensional point cloud and camera pose information to reconstruct a three-dimensional scene by using a 3DGS method, and generating a three-dimensional Gaussian kernel to represent an input scene; S3. introducing a Sobel edge detector, and respectively extracting edges of transparent objects in a rendered image generated by the three-dimensional Gaussian kernel and a real image, and calculating an edge loss function of a rendered edge map and a real edge map; S4. combining the edge loss function and an original 3DGS loss function to jointly optimize a three-dimensional scene reconstruction model; and S5. overall optimizing the scene and outputting a final three-dimensional reconstruction result. The application aims to improve the reconstruction quality of a 3DGS model, and further improve the three-dimensional reconstruction effect of transparent objects.

The present invention provides a low-power receiver for MIPI C-PHY, comprising: an oscillator, a delay cell calibration, an edge detection and pulse generator, and a sampler & wire state decoder. The delay cell calibration is connected to the oscillator signal and receives the frequency signal generated by the oscillator. The edge detection and pulse generator is signal-connected to the delay cell calibration and is used to receive signals transmitted by the MIPI C-PHY. The sampler and wire state decoder is signal-connected to the edge detection and pulse generator and is used to receive signals transmitted by the MIPI C-PHY and output a decoding result. The invention utilizes the known oscillator pulse width (OSC clock high period width, i.e., the duration of the high level of the pulse signal), which is the delay time of one delay cell. After passing through the delay cell calibration circuit, the number of delay cells within the selected oscillator OSC cycle time can be determined. With the known data rate, a selection code is sent to the same delay cell in the edge detector to obtain the delay time for edge pulse generation. The edge pulse is sent to the sampler for data sampling, and after sampling, the state of the decoding action is determined.

The invention discloses a tunnel surrounding rock crack detection and direction prediction method based on an improved YOLO model. The method comprises the following steps: constructing a crack detection model based on a YOLOv11 architecture; inputting a to-be-detected image into the fracture detection model, and identifying a candidate fracture region by the fracture detection model; extracting a corresponding image block from the candidate crack region, converting the image block into a grayscale image, and extracting edge features by using an edge detector to obtain an edge image; geometric primitives are extracted from the edge graph through probability Hough transform, and a line segment candidate set is generated; and selecting the longest line segment from the line segment candidate set to determine the main direction of the fracture, and performing direction classification based on the main direction. According to the method, the detection capability of small-size and irregular-shaped fractures is effectively improved, a geometric analysis module based on probability Hough transform is introduced, and the principal direction of the fractures is accurately estimated on the basis of detection. According to the end-to-end frame, the detection precision is remarkably improved, reliable direction estimation is provided, and the application potential of the end-to-end frame in actual structure monitoring is enhanced.

A metrology system for determining a characteristic of a structure on a substrate. The metrology system comprises an illumination source to output radiation to an objective. The objective is configured to direct radiation toward the structure on the substrate and direct radiation diffracted from the structure toward a detector. The detector comprises an event based camera to detect changes in intensity of radiation incident on the event based camera, and outputs an event data stream conveying the detected events. The detector comprises an edge detector o receive the event data stream and to output movement detection data representing a movement of the structure on the substrate relative to the objective or camera. The detector comprises a processor to receive the output movement detection data and determine the characteristic of the structure on the substrate based on the movement detection data and 0 properties of the structure on the substrate.

The invention discloses a visible light infrared target tracking method and system based on edge information dynamic enhancement, and mainly solves the problems of insufficient utilization of target structure information and equal utilization of modal information in the prior art. According to the scheme, the method comprises the following steps: intercepting a multi-mode video into multiple frames of images according to time, and preprocessing by taking a GTOT data set as a training set; a tracking model comprising a local gradient edge detector, a hybrid multi-expert dynamic selection unit, an edge adapter and a backbone network is constructed, and the backbone network is in parallel connection with the local gradient edge detector, the edge adapter and the hybrid multi-expert dynamic selection unit which are connected in series and is used for extracting multi-modal image features; inputting the preprocessed training set into a tracking model for iterative training; and inputting the test set RGBT234 into the trained tracking model to output the position information of the target in each frame of the video. The method can efficiently utilize the edge information of the target to enhance the representation capability of the whole network to the target, improves the tracking performance, and can be used for infrared warning, electric power, medical treatment and the like.

The invention discloses a large model auxiliary damage detection system and method based on verification prompt and reflux distillation, and belongs to the field of machine vision and industrial detection. The system comprises an edge detector, an adaptive verification distributor, a verification prompt generator, a cloud verifier, a reflux distillation module and a resource and physical monitoring subsystem. Preliminarily detecting an output result, uncertainty and characteristics at the edge end; the distributor is combined with uncertainty, physical parameters and resource states to trigger a cloud rechecking and uploading mode; generating a compressed structured verification prompt data packet during triggering; the cloud rechecks through a large model and integrates physical consistency verification; the reflux distillation module utilizes cloud feedback to finely adjust the edge model; and the monitoring subsystem dynamically adjusts the strategy. The verification prompt compression enables the single-time uploading data volume to be reduced; the detection accuracy is improved through end-cloud cooperation and physical verification, and false positive is inhibited; the adaptive strategy balances the delay cost and the accuracy of key events; and the edge model is continuously and adaptively optimized through reflux distillation.

A propagating light field detection system and method are disclosed. The detection system includes a housing, a linear displacement stage, a knife-edge detector, a light-transmitting aperture, and a control center. Both the linear displacement stage and the knife-edge detector are mounted inside the housing, and the light-transmitting aperture is located on the side of the housing. The linear displacement stage is used for high-precision scanning along the beam propagation direction, and the knife-edge detector is used to scan the light field distribution across the transverse cross-section of the beam. Using the detection system and method of this invention, the spatial propagation characteristics of a laser light field can be accurately measured, and the depth of focus and focal point size can be detected, demonstrating significant potential application value in laser processing and laser optics.

The invention relates to a well logging technology, and discloses a multilayer stratum inversion method of a while-drilling electromagnetic edge detector, which comprises two stages of offline training and online inversion. In the offline stage, firstly, multiple stratigraphic structure types are constructed, and training samples are generated; performing feature enhancement and reconstruction on the forward modeling original response vector of the sample to improve the data dimension and sensitivity; then, for each stratum type, training through a supervised descent method to obtain a special gradient optimization set; in the online stage, the same feature reconstruction is carried out on the original response measured in real time, the stratum type to which the original response belongs is intelligently judged, and the matched gradient optimization set is automatically called to carry out rapid iterative inversion. Dynamic attenuation regularization and geological prior constraint are fused in inversion iteration, noise is effectively suppressed, and stability and geological rationality are guaranteed. According to the method, rapid, accurate and stable inversion of a complex multi-layer stratum is realized, and the reliability of geosteering while drilling is remarkably improved.

The application discloses a kind of unmanned aerial vehicle hyperspectral push-broom imaging geometric correction methods, comprising the following steps: step 1, according to POS data, DEM data, using collinear condition equation obtains orthophoto, using Canny / Devernay subpixel edge detector, extract the edge point of corrected orthophoto;Step 2, using BM3D algorithm and median filter smooth orthophoto obtained in step 1, the image that will eliminate sawtooth distortion is used as reference image, using Canny / Devernay subpixel edge detector, extract the edge point of this image;Step 3, to the orthophoto obtained in step 1, select the area needing further correction, using the edge point extracted in step 1, 2 calculates the roll angle correction number, and corrects roll angle error;Step 4, the roll angle in POS data is replaced by the roll angle corrected in step 3, repeat step 1, and obtain the orthophoto that eliminates the distortion of roll angle error.The application can effectively correct the distortion of unmanned aerial vehicle line scanning hyperspectral image.

In an identification information receiving device that receives identification information defined by a plurality of bits and superimposed on a current or a voltage of a power line by an identification information transmitting device that is connected to the identification information receiving device via the power line, an edge detector detects an edge of an input signal that indicates a voltage value corresponding to the current or the voltage of the power line. A T flip-flop circuit inverts an output, each time the edge of the input signal is detected.

This invention discloses a design method and a metalens device for monolithic dual-mode edge detection, belonging to the field of optical edge detection technology. When used in conjunction with a linear polarizer, this device can achieve isotropic two-dimensional and directionally adjustable one-dimensional edge detection according to the polarization state of the incident light. The phase distribution of the nanoarray in the device in the x and y polarization directions is the bright-field imaging phase and the phase difference between the phases, respectively. Each nanounit rotates counterclockwise with its lattice center as the rotation center. The rotation angle relative to the initial posture (when the angle between its major axis and the x-axis is 0° or 90°) is half the angle between the line connecting the lattice center and the origin of the x-y coordinate system and the x-axis. Based on this, this invention can achieve edge detection imaging with a wide working bandwidth, high edge detection resolution, and convenient mode switching without the need for other imaging devices.

This invention discloses a method and apparatus for detecting lesions in X-ray images. The method includes: an X-ray image feature extraction module for extracting image features from X-ray images; an edge-guided spinal feature enhancement module for extracting spinal edge maps from X-ray images using an edge detector and enhancing spinal features using the spinal edge maps; a candidate feature extraction module for obtaining candidate features from the enhanced spinal features; a context-aggregated structural relationship reasoning module for obtaining query features and context features from the enhanced spinal features, mining the correlation between the query features and context features to obtain aggregated features, and using the aggregated features to compensate for the candidate features; and a spinal lesion detection head for detecting spinal lesions using the compensated candidate features. The apparatus includes: a processor and a memory. This invention delves into the characteristics of spinal lesions in X-ray images to achieve more accurate lesion perception, thereby improving the performance of spinal lesion detection.

The application provides a polarization-independent BIC metal super surface optical edge detection device and a design method, the method shows a simple and easy-to-process metal super surface, the device can realize high-quality polarization-independent two-dimensional optical edge detection with high numerical aperture and high efficiency. The design scheme provided by the method takes silicon dioxide as a substrate and C4 symmetric L-shaped aluminum metal patch as a basic unit, adjusts and controls the spatial dispersion of BIC by fine-tuning the perturbation parameters of the super surface, and realizes the polarization-independent second-order spatial differential operation. The design scheme works at a frequency of 0.6718 THz, the numerical aperture (NA) can be as high as 0.56, the peak efficiency can be as high as 36%, and the high-quality edge detection performance can be maintained. The work can realize compact, low-power and ultrafast optical edge detection, and shows great potential in various applications such as optical imaging and computer vision.