131 results about "Acutance" patented technology

The invention provides an intelligent control method for a floodlight illumination system, and the method comprises the steps: obtaining an environment surface roughness coefficient, spectral reflection distribution and spatial geometric change data, determining the wall surface material absorptivity and shielding region coordinates, and obtaining environment reflection characteristics and shielding distribution data; analyzing shielding boundary sharpness and shielding angle distribution according to the spatial geometric change and the coordinates of the shielding area, calculating a luminous flux distribution curve of a local dark area, and judging whether the influence range of the dark area exceeds a preset threshold range or not; calculating the illumination coverage integrity according to the updated light field distribution and polarization reflection characteristics, determining the boundary sharpness of a light beam, and obtaining a light field uniformity parameter; and according to the light field uniformity parameter and the shielding thermal radiation effect, calculating and optimizing light field parameters.

The invention relates to the technical field of industrial vision, in particular to an aviation part crack detection and repair method which comprises the following steps: acquiring an aviation part optical image at a reference time point and an aviation part optical image at a to-be-detected time point; according to the method, logarithmic polar coordinate transformation is carried out on different time point images, rotation, scaling and translation parameters are extracted, an image registration relation is established, the structural consistency of time sequence images in a local area is enhanced, a structural tensor is constructed for each pixel neighborhood, the change characteristics of the dual-time-phase tensor are compared, a structural change saliency map is generated, and the structural change saliency map is obtained. Sensitive capture of a tiny deformation area is achieved, the responsiveness to an initial crack is improved, then a crack propagation interval is further refined into a main crack path in a self-adaptive threshold segmentation and skeleton extraction mode, the tip acutance of the crack is calculated in combination with tip contour information of a geometric boundary of the crack, and the initial crack is obtained. And quantitative support is provided for the crack danger degree.

An automated resolution and sharpness calibration system for target scanning includes an imaging reference located on a scanner's scanning stage and an optical sensor configured to capture reference images before and after scanning a target. The system comprises a processor and a memory with instructions that enable the processor to receive the captured reference images, determine image metrics based on reference features, and detect deviations in these metrics between the images. The processor classifies the detected deviation to identify its cause and generates an alert signal accordingly. This system ensures precise calibration by analyzing deviations in image metrics, thereby maintaining optimal scanning performance and accuracy.

The invention relates to the technical field of image data processing, in particular to a water-soluble fertilizer dissolution rate visual measurement method and system, and the method comprises the steps: obtaining a sequence image of a dissolution process; performing multi-scale frequency domain analysis on the image, and calculating a scale convergence index by using the deviation between a local phase and a weighted average phase; a spectral sharpness factor is calculated according to the ratio of the high-frequency energy response to the low-frequency energy response; fusing the scale convergence index and the spectral sharpness factor to obtain a solid boundary probability; segmenting the particle region based on the probability and converting the particle region into physical mass, and differentiating a mass change sequence to obtain a dissolution rate. By fusing the phase consistency and the frequency domain energy distribution characteristics, the physical entity edge and the optical refraction halo are effectively distinguished, the refraction interference problem during high-concentration dissolution is solved, and the dissolution rate measurement accuracy is improved.

The invention relates to the technical field of deep learning model training, solves the technical problem that gradient pointing is inaccurate when model parameters are updated in a traditional SAM algorithm, and particularly relates to an image classification method based on sharpness perception minimization. A gradient direction correction mechanism is introduced for an image classification task, so that the stability in the optimization process and the generalization ability of the model on image recognition test data are remarkably improved. According to the method, the gradient disturbed by the SAM algorithm is corrected by using the feature vector, and the abnormal component of the gradient in the feature vector direction is effectively reduced, so that model parameter updating is prevented from pointing to a sharp region of a loss function. The correction of the optimized path significantly improves the accuracy of gradient updating in the training process of the image classification model, so that the model can learn more discriminative visual feature representation, and finally the classification accuracy and the generalization performance of the model on a test set are improved.

The invention relates to the technical field of face recognition, and discloses a face recognition system and a detection method thereof, and the recognition system is composed of a face image acquisition module, an environment perception and preprocessing module, a face detection feature extraction module, a comprehensive analysis and comparison module and a data storage monitoring module. According to the method, environmental factors such as illumination intensity, light source type and color temperature are detected through a sensor, corresponding preprocessing is carried out on the image, the influence of the environmental factors on the recognition precision can be reduced by correcting illumination unevenness, denoising and enhancing contrast and sharpness, meanwhile, image compensation and denoising are carried out, illumination unevenness is compensated, and denoising processing is carried out, so that the recognition precision is improved. The beneficial effects of comprehensive analysis of age, health conditions and illumination condition environmental factors through face recognition and more stable recognition precision are achieved.

The invention discloses an X-ray image artifact definition automatic correction and enhancement method based on deep learning, and belongs to the technical field of image artifact correction, and the method specifically comprises the steps: obtaining an X-ray image, employing frequency domain decomposition and edge detection to generate an artifact image and a sharpness image, and extracting an imaging parameter set and an anatomical region label; establishing a parameter-driven artifact migration network, inputting an imaging parameter set and an artifact graph, learning an artifact vector field and phase prior, and forming reversible representation of an artifact source; constructing a double-branch decoder, dissecting branches to generate a structural skeleton diagram, and performing cross attention fusion to obtain a candidate correction field; according to the candidate correction field, geometric remapping and sub-band compensation are carried out on the image to obtain an intermediate correction result image, and an artifact vector field is sent to a feedback loop; and the consistency of the sharpness graph and the structural skeleton graph is taken as a loss item, artifact residual constraint is combined, a correction field and decoder parameters are optimized, and non-anatomical textures are suppressed.

The invention provides a sweet menu embryo pelleting seed breakage rate detection method based on machine vision, and relates to the technical field of seed detection.The method comprises the steps that firstly, visible light and near-infrared images of seeds under standard illumination and non-standard visible light images under multiple light sources are collected, and reflectivity calibration and polarization difference correction processing are carried out; features such as colors and textures are extracted based on visible light and near-infrared reflectivity images, and multispectral feature vectors are constructed; a surface normal vector is calculated by using a photometric stereoscopic vision technology and combining a plurality of non-standard visible light images, an initial three-dimensional height map is generated through integration, and a plurality of abnormal judgment features are extracted to mark a suspected damaged area. And further calculating an edge sharpness index, removing a mechanical damage area, and determining a final damage area and a three-dimensional contour thereof. And finally, combining the volume loss rate, the projection area ratio and other multi-dimensional features with the multispectral feature vector to construct a damage probability prediction model, thereby realizing accurate and efficient detection of the seed damage rate.

The invention discloses an atherosclerotic plaque risk assessment method and system based on artificial intelligence, relates to the technical field of medical image analysis, and provides the following scheme: synchronously acquiring medical image sequence and physiological signal time sequence data, generating a synchronization data set through time-space registration, and performing time-space registration on the synchronization data set; and inputting the synchronized data set into a pre-trained LSTM neural network, and outputting a dynamic artifact offset field which represents a displacement vector of a medical image pixel point along with a physiological signal phase. Gradient amplitude analysis is performed on the medical image sequence frame by frame to generate the gradient characteristic evolution graph, high, medium and low response areas are divided according to the dynamic gradient threshold interval, and the displacement correction vectors are configured for the areas, so that the problem of gradient response heterogeneity compensation deficiency in a conventional correction strategy is solved, adaptive deformation correction of an anatomical structure is realized, and the correction accuracy is improved. And low-gradient region structure dispersion and vascular wall edge sharpness attenuation caused by global compensation are inhibited.

The invention discloses a fingerprint verification method and device based on double-finger fingerprint image acquisition, and relates to the technical field of biological recognition, and the method comprises the following steps: deploying a light sensor, carrying out the primary adjustment of an illumination light source, collecting the light intensity data of the fingerprint surface when two fingers press through the light sensor, and obtaining a fingerprint verification result according to the light intensity data; the method comprises the following steps: deploying a light sensor, constructing a closed-loop adjustment mechanism of primary adjustment, secondary adjustment and quality feedback, acquiring light intensity data of the fingerprint surfaces of the two fingers in real time, judging an abnormal condition, performing secondary adjustment on an illumination light source, acquiring fingerprint images of the two fingers, and judging whether the quality of the fingerprint images reaches the standard or not. According to the two-finger fingerprint verification method, the abnormity of insufficient light source intensity or improper angle is pointedly corrected through secondary adjustment, and the image quality evaluation standard combining the edge sharpness and the effective area proportion provides reliable data support for two-finger fingerprint verification, and the fingerprint verification accuracy and the actual application effect are improved.

The invention discloses a near-infrared confocal microscopic imaging system and a near-infrared confocal microscopic imaging method, and relates to the technical field of optical microscopic imaging. The method comprises the following steps that near-infrared exciting light sequentially passes through a beam expanding system, a liquid lens and a confocal scanning light path to irradiate a to-be-detected sample so as to generate a near-infrared fluorescence signal, and the liquid lens is used for adaptively adjusting the focal length and wavefront morphology according to the depth of the to-be-detected sample; a photomultiplier detects the near-infrared fluorescence signal after non-focus scattered light is filtered out through the confocal scanning light path and the pinhole, and a fluorescence signal image is obtained; and according to the fluorescence signal image, carrying out iterative optimization on the focal length and phase modulation parameters of the liquid lens to maximize the image sharpness or the signal-to-noise ratio and realize the three-dimensional microscopic imaging of the deep tissue. According to the near-infrared confocal microscopic imaging method provided by the invention, deep near-infrared microscopic imaging with low phototoxicity, low cost and high resolution can be realized.

The invention provides an intelligent feedback control glass outer diameter screening method, which comprises the following steps: collecting the reflected light intensity and refracted light angle of the surface of glass, analyzing the spatial frequency distribution characteristics of local scattering intensity, identifying the surface microcrack distribution mode and quantifying the defect depth change; distinguishing real structure defects and surface tiny scratches, and obtaining defect boundary sharpness and defect size distribution parameters; based on the denoised signal data, dynamic evolution characteristics of defect distribution are extracted, and a space-time incidence matrix containing light intensity spatial frequency periodicity and angle deviation periodicity is constructed; carrying out weight redistribution on the space-time incidence matrix by utilizing a dynamic weight distribution coefficient, and carrying out weighted fusion on local scattering intensity information and refracted light phase difference information to generate a fused feature signal; based on the space-time characteristic parameters of the fused characteristic signals, the influence of sudden drop of reflected light intensity and signal distortion on outer diameter judgment is reduced to the maximum extent, and optimized outer diameter judgment signals are output.

The invention belongs to the technical field of image processing, and relates to an image video rain removal method based on local perception modeling and multi-stage span fusion, which adopts parallel branch networks with three level scales to respectively and sequentially carry out feature extraction, feature extraction and multi-stage span fusion on raining images with corresponding resolution sizes, performing high-level semantic information mining on the input raining image to obtain a high-dimensional feature map; high-resolution image reconstruction: adopting a local set perception implicit representation module, mapping the high-dimensional feature map to a continuous RGB color space through linear coordinate projection, and compensating continuous textures and edges sheltered by rain in the image to obtain a reconstructed image; and cross-scale feature fusion: carrying out information fusion between stages and scales on the image through double branches, and aligning and enhancing multi-scale semantics to obtain a high-resolution rain-removed image. According to the method, artifacts of a raining image can be suppressed in a multi-domain and multi-scale level, and sharpness and consistency are enhanced.

The invention relates to the technical field of digital image processing, in particular to a knitted label scanning image completion operation method, which comprises the following steps of: acquiring a knitted label image and converting the knitted label image to a frequency domain, locking a texture fundamental frequency based on energy distribution, executing protective denoising, repairing a texture phase flow field, extracting a character stroke skeleton in combination with gradient analysis, and completing the completion of the knitted label scanning image. According to the method, the image is converted to the frequency domain, the texture fundamental frequency energy is locked, the periodic structure characteristics of the knitted fabric are separated and protected, continuous solution is carried out on the phase field, smooth reconstruction of the warp and weft texture manifold is achieved, the phase dislocation defect caused by local pixel matching is eliminated, and the method has the advantages of being high in accuracy and high in reliability. A comprehensive steering vector field and a character stroke skeleton path are generated through orthogonal gradient synthesis, strong geometric constraint is introduced in the pixel filling process, it is ensured that yarn textures follow the original weaving trend after repairing, and character stroke coherence and edge sharpness are ensured.

The invention discloses an unmanned aerial vehicle-mounted through-wall SAR motion error inversion method based on image sharpness and a GCBP algorithm. In order to solve the problem that SAR images are seriously defocused due to airborne platform motion errors, a traditional self-focusing method is popularized to the field of unmanned aerial vehicle-mounted through-wall SAR, and an accurate unmanned aerial vehicle-mounted through-wall SAR signal model is constructed based on the refraction effect of a wall; motion errors are analyzed from geometric characteristics, frequency dependence characteristics, time-varying characteristics and space-varying characteristics, and an accurate phase error model is constructed by adopting sub-aperture division and sub-band division strategies in combination with a GCBP algorithm; and finally, based on an image sharpness criterion and a motion continuity constraint, realizing accurate estimation of the motion error of the unmanned aerial vehicle-mounted through-wall SAR by adopting a gradient descent method combined with a line search strategy. The actual measurement experiment verifies that the method has good refocusing effect and parameter inversion performance in the field of unmanned aerial vehicle-mounted through-wall SAR (Synthetic Aperture Radar).

A method includes capturing images of a surface of a workpiece using an optical sensor. Each image respectively images a region of the surface. Each image is assigned a defined 6-DOF pose of the optical sensor relative to the workpiece and a defined focal plane position of the optical sensor. The captured images form an image stack. The method includes determining a sharpness value for each picture element of each image of the image stack to generate a sharpness map for each image. The sharpness maps of the images form a sharpness map stack. The method includes transforming the sharpness maps of the sharpness map stack into a defined reference system based on the respective assigned 6-DOF pose and focal plane position of the optical sensor in order to generate a sharpness cloud. The method includes generating a surface profile of the workpiece based on the sharpness cloud.

The application discloses a kind of for promoting image processing method under glare, shadow environment, comprising: using dynamic compensation algorithm to focus point offset amount is adjusted in real time, by monitoring the change trend of contrast value and the convergence state of definition index in focusing process, the correctness of current focusing direction and the suitability of focusing amplitude are judged;The transition characteristics of shadow and non-shadow area are analyzed by shadow edge enhancement algorithm, the direction and amplitude of focus point transformation are dynamically adjusted, and the real position locking result of shadow edge is obtained;For the edge sharpness distribution parameter after promotion, the focus consistency of overall image is verified using feature matching algorithm, and the evaluation result of global focus quality is obtained by calculating the definition difference between different regions.

An information processing device includes an image transformation unit, a left-right difference estimation unit, and an image generation unit. The image transformation unit performs warping to move positions of a feature point of a right-eye image and a feature point of a left-eye image based on right-eye and left-eye viewpoint information. The left-right difference estimation unit estimates a portion where a difference exceeding an allowable level occurs, due to the warping, between the right-eye image and the left-eye image as an inconsistent portion. The image generation unit makes a sharpness of the inconsistent portion different between the right-eye image and the left-eye image.

This invention discloses an automatic calibration method for airborne lidar placement deviation based on voxel sharpness. The method includes: selecting a natural calibration site containing building facades, linear markers, and point targets; collecting point cloud and POS data via a UAV flying in multiple headings; modeling the placement deviation to be calibrated using six parameters; optimizing the parameters using a genetic algorithm; using voxel sharpness as the fitness function; voxelizing the point cloud in the region of interest; counting the number of points within each voxel and calculating the sharpness score; a higher sharpness score indicates a more pronounced clustering of point clouds in the feature region; and iteratively optimizing the method through selection, crossover, and mutation operations to output the optimal placement deviation parameters. This invention uses an ordinary basketball court as a natural calibration site, eliminating the need for artificial targets. After calibration, the point cloud sharpness is improved by more than 150%, and the positioning accuracy reaches the centimeter level.

The invention relates to the technical field of microscopic imaging, and discloses a multi-level physical model and double-domain deep learning structured light illumination imaging reconstruction method, which comprises the following steps: firstly, constructing a multi-level physical degradation model based on illumination light field modulation, optical diffraction sensor discrete sampling and photoelectric noise, and generating a simulation data set containing random errors; then constructing a deep neural network containing spatial domain and frequency domain branches, and performing iterative training on the network by using a joint loss function; and finally, acquiring an original structured light stripe image of a real sample, inputting the original structured light stripe image into the trained model to execute end-to-end reasoning, and outputting a super-resolution fluorescent microscopic image. According to the method, the problem of obtaining real pairing training data is solved through physical simulation, high-frequency detail recovery is enhanced by utilizing frequency domain constraint, reconstruction artifacts are effectively removed, the imaging resolution is improved, and the spatial resolution and edge sharpness of a reconstructed image are improved while the ringing effect and background noise are effectively suppressed.

The invention relates to the technical field of image processing, and discloses a galvanized wire surface zinc layer defect identification method and system. The method comprises the steps of collecting a galvanized wire surface image to extract a boundary gradient sequence, performing median filtering and abnormal point elimination on the gradient sequence, verifying boundary sharpness consistency under multiple scales, judging defect types by combining frequency domain analysis and gray features, and tracking defect time sequence evolution features to calculate a quality risk score and output a grading result. According to the method, the accuracy and comprehensiveness of quantitative evaluation of the defect severity are improved.

Various implementations provide unvarnished video based on adjusting camera parameters according to environmental modeling. Environmental characteristics may be determined based on modeling a physical environment from sensor data captured via one or more sensors. For example, this may involve determining ambient light source optical characteristics, ambient surface optical characteristics, 3D mapping of the environment, user behavior, prediction of optical characteristics of light entering the camera, and the like. The method may involve determining camera parameters of an image captured via an image sensor based on environmental characteristics. For example, the method may determine exposure, gain, tone mapping, color balance, noise reduction, sharpness enhancement. The method may determine camera parameters based on user information (e.g., user preferences, user activity, etc.). The method may involve providing a transparent video of the physical environment based on the determined camera parameters.

The invention discloses an image domain azimuth multichannel SAR (Synthetic Aperture Radar) phase error estimation method and device. The method comprises the following steps: acquiring a multichannel SAR image; performing multi-round iterative optimization on the phase error estimation vector of the multi-channel SAR until a convergence condition is reached, and obtaining an optimal phase error estimation value; wherein the optimal phase error estimation value is a phase error estimation vector with the best fitness in the last round of iterative optimization, and the fitness is calculated according to the multi-channel SAR image and the phase error estimation vector and is used for representing the sharpness of the multi-channel SAR image reconstructed based on the phase error estimation vector. The method is not limited by the number of channels during estimation, and the estimation precision is good.

A method for analyzing an image, called “analysis image”, of a dental arch of a patient. The method in which the analysis image is submitted to a neural network, in order to determine at least one value of an image attribute relating to the analysis image. The analysis image is a photograph or an image taken from a film. The image attribute relates to a position or an orientation or a calibration of an acquisition apparatus used to acquire the analysis image or a combination thereof, or a quality of the analysis image, and in particular relating to the brightness, to the contrast or to the sharpness of the analysis image, or a combination thereof.

The invention provides an image amplification method for gradient adaptive interpolation and noise suppression, and belongs to the field of image processing before ink-jet printing. Extracting a sawtooth texture position of the gradient mask image as a noise point through a directional non-extreme suppression operator; performing median filtering on the noise points to generate an intermediate stage gradient image; self-adaptive interpolation processing is carried out according to the edge texture position of the image in the middle gradient stage; the interpolation processing result is corrected through gradient weight and distance weighting, and interference fringes are suppressed; and sending the amplified image block after interference removal into a buffer area. According to the image amplification method based on gradient adaptive interpolation and noise suppression, the processing efficiency is remarkably improved through intelligent blocking and parallel optimization, and sawtooth noise caused by high-power amplification is accurately eliminated by combining a hierarchical amplification strategy and a directional non-extreme suppression operator. And the adaptive interpolation algorithm based on gradient guidance effectively inhibits interference fringes while keeping edge sharpness, and realizes synchronous optimization of operation speed and visual quality.

The present application relates to the technical field of digital image processing, in particular to a kind of knitted label scanning image completion operation method, comprising the following steps: acquisition knitted label image and conversion to frequency domain, based on energy distribution locking texture base frequency, execute protective denoising, repair texture phase flow field, extract text stroke skeleton in conjunction with gradient analysis, fill in missing pixels, output knitted label completion image, in the present application, by transforming image to frequency domain and locking texture base frequency energy, separate and protect knitted fabric periodic structure characteristics, utilize to phase field is carried out continuity solution, realize warp and weft texture manifold smooth reconstruction, eliminate the phase misplacement drawbacks caused by local pixel matching, combined with orthogonal gradient synthesis generates comprehensive guiding vector field and text stroke skeleton path, introduce strong geometric constraint in the process of pixel filling, ensure that the yarn texture after repair follows original weaving direction, and ensure text stroke coherence and edge sharpness.

The application provides a RAW domain low-illumination image enhancement method and system, relates to the technical field of image processing, and can effectively avoid amplification of noise in the subsequent enhancement process, reduce noise and color distortion of the RAW domain enhanced image by separating the content signal graph and the noise residual graph of the RAW domain low-illumination image. Through the content signal graph, the light rays of the camera light path are simulated and light integration is carried out, so that the physical level detail reconstruction can be carried out on the region with serious structure loss in the RAW domain low-illumination image, and the global light consistency and edge sharpness better than the traditional two-dimensional convolution network are realized. By introducing the target double-flow recurrent neural network, the optimization conflict between denoising and color restoration in the low-illumination enhancement can be solved, the progressive collaborative improvement of the image quality can be realized, and the high harmony of the final enhancement result in the three dimensions of brightness, detail and color is ensured.

A method is for determining a sharpness condition of a saw chain of a chain saw. The saw chain includes at least one cutting link having an upper cutting blade. An upper cutting blade image of the upper cutting blade of the cutting link is recorded using an imaging device. An evaluation of the upper cutting blade image is performed using an evaluation unit, which includes an artificial neural network. A sharp condition and a dull condition of the saw chain are defined in the evaluation unit. The saw chain is assigned on the basis of the upper cutting blade image using the artificial neural network to the sharp condition or the dull condition.

The invention relates to the technical field of image processing, and discloses an image scaling method and device, equipment and a medium, and the method comprises the steps: respectively obtaining the height information and width information of an initial image and a target image, determining a longitudinal scaling based on the height information, and determining a transverse scaling based on the width information; mapping a target pixel position in the target image to a floating point coordinate position in the initial image based on the transverse scaling and the longitudinal scaling respectively to obtain a corresponding transverse floating point coordinate and a corresponding longitudinal floating point coordinate; determining a pixel mean value in a preset pixel position range in the initial image according to the transverse floating point coordinates and the longitudinal floating point coordinates, and obtaining a neighborhood pixel mean value; and limiting an interpolation weight range according to a preset sharpness adjustment parameter, and performing linear interpolation on the neighborhood pixel mean value to obtain a target pixel of the target image so as to zoom the initial image. The invention aims to improve the visual quality and detail retention effect of the zoomed image.

The invention discloses a video monocular depth prediction method and system based on multilevel three-dimensional geometric constraints, and the method comprises the steps: firstly constructing a monocular depth prediction network and a pose prediction network, and then carrying out the prediction of a sample video through employing the monocular depth prediction network and the pose prediction network, and obtaining an image prediction depth and a pose prediction value. Respectively constructing a target view three-dimensional point cloud and a source view three-dimensional point cloud according to the image prediction depths of the target view and the source view; according to the method, the function of carrying out accurate and efficient monocular depth prediction on the video by adopting multi-level explicit three-dimensional geometric constraints is realized, a deep 3D geometric relationship can be converted into an optimizable loss function, and image information in a sequence is fully utilized, so that an accurate, smooth, continuous and robust depth map is output; and the performance of depth prediction in the aspects of edge sharpness, surface smoothness and cross-view consistency is remarkably improved.