517 results about "Deblurring" patented technology

A method for deblurring an image. A first image is captured. A second image is captured, wherein the second image is more blurred and more exposed than the first image. The second image is deblurred based on the first image.

Systems and methods for providing a substantially de-blurred image of a scene from a motion blurred image of the scene are disclosed. An exemplary system includes a primary detector for sensing the motion blurred image and generating primary image information representing the blurred image, a secondary detector for sensing two or more secondary images of the scene and for generating secondary image information representing the two or more secondary images, and a processor for determining motion information from the secondary image information, estimating a point spread function for the motion blurred image from the motion information, and applying the estimated point spread function to the primary image information to generate information representing the substantially de-blurred image.

Systems and methods for providing a substantially de-blurred image of a scene from a motion blurred image of the scene are disclosed. An exemplary system includes a primary detector for sensing the motion blurred image and generating primary image information representing the blurred image, a secondary detector for sensing two or more secondary images of the scene and for generating secondary image information representing the two or more secondary images, and a processor for determining motion information from the secondary image information, estimating a point spread function for the motion blurred image from the motion information, and applying the estimated point spread function to the primary image information to generate information representing the substantially de-blurred image.

Input data comprising a video signal is processed using a combination of a known image processing method to deblur, or sharpen, the image and convolution with Pixon™ kernels for smoothing. The smoothing process utilizes a plurality of different size Pixon™ kernels which operate in parallel so that the input data are convolved with each different Pixon™ kernel simultaneously. The smoothed image is convolved with the point response function (PRF) to form data models that are compared against the input data, then the broadest Pixon™ kernel that fits the input data within a predetermined criterion are selected to form a Pixon™ map. The data are smoothed and assembled according to the Pixon™ map, then are deconvolved and output to a video display or other appropriate device, providing a clearer image with less noise.

The present Invention relates to a camera system suitable for use in minimally invasive surgery (MIS), among other applications. In at least one embodiment, the camera system includes an autonomous miniature camera, a light source assembly providing features such as steerable illumination and a variable radiation angle, and a control and processing unit for processing images acquired by the camera Io generate improved images having reduced blurring using a deblurring algorithm.

Systems and methods are disclosed for deblurring a captured image using parametric deconvolution, instead of a blind, non-parametric deconvolution, by incorporating physical constraints derived from sensor inputs, such as a motion sensor, into the deconvolution process to constrain modifications to the point spread function. In an embodiment, a captured image is deblurred using a point spread function obtained from the cross-validation of information across a plurality of image blocks taken from the capture image, which image blocks are deconvolved using parametric deconvolution to constrain modifications to the point spread function.

The present disclosure provides for improved image analysis via novel deblurring and segmentation techniques of image data. These techniques advantageously account for and incorporate segmentation of biological analytes into a deblurring process for an image. Thus, the deblurring of the image may advantageously be optimized for enabling identification and quantitative analysis of one or more biological analytes based on underlying biological models for those analytes. The techniques described herein provide for significant improvements in the image deblurring and segmentation process which reduces signal noise and improves the accuracy of the image. In particular, the system and methods described herein advantageously utilize unique optimization and tissue characteristics image models which are informed by the underlying biology being analyzed, (for example by a biological model for the analytes). This provides for targeted deblurring and segmentation which is optimized for the applied image analytics.

A method for prediction of an indication of motion using input from an event-based camera includes receiving events captured by an event-based camera, wherein each of the events represents a location of a change in pixel intensity, a polarity of the change, and a time. The method further includes discretizing the events into time discretized event volumes, each of which contain events that occur within a specified time range. The method further includes providing the time discretized event volumes as input to an encoder-decoder neural network trained to predict an indication of motion using a loss function that measures quality of image deblurring; generating, using the neural network, a prediction of the indication of motion. The method further includes using the prediction of the indication of motion in a machine vision application.

The invention belongs to the technical field of computer digital image processing, and particularly relates to an image deblurring method based on an aggregation expansion convolutional network. The method comprises the steps of constructing a deep neural network, generating a network based on a condition countermeasure, wherein the network comprises a generator and a discriminator, the generatorstructure uses a stacked self-encoder module, and the self-encoder module is connected with a jump through a self-encoder structure, a residual error module is used on a construction module, residualerror module uses a residual network and multi-channel aggregation expansion convolution, and the discriminator uses a five-layer convolutional neural network; training the deep neural network: usingfuzzy image data set in public and real scenes, using image content loss function and a countermeasure loss function to train the deep neural network constructed in the previous step, and using a trained network model to carry out deblurring processing on a blurred image. According to the method disclosed by the invention, the deblurring effect can be ensured, a blurred image can be quickly and efficiently restored to a clear image, and the image deblurring efficiency can be greatly improved.

A method and device for deblurring an image having pixels. A blurred image is downloaded into a systolic array processor having an array of processing logic blocks such that each pixel arrives in a respective processing logic block of one pixel or small groups of pixels. Data is sequentially exchanged between processing logic blocks by interconnecting each processing logic block with a predefined number of adjacent processing logic blocks, followed by uploading the deblurred image. The processing logic blocks provide an iterative update of the blurred image by (i) providing feedback of the blurred image prediction error using the deblurred image and (ii) providing feedback of the past deblurred image estimate. The iterative update is implemented in the processing logic blocks by u(n+1)=u(n)−K*(H*u(n)−yb)−S*u(n).

The invention provides a method of license plate image deblurring based on depth learning. The method is divided into a data set preprocessing stage, a training stage and a testing stage. In the dataset preprocessing stage, the license plate region in the image is determined, the license plate characters are segmented and the image size is normalized, Gaussian noise is added to obtain the training set. In the training phase, the linear sum of the mean square error, the gradient error and the discrimination error of the restoration result is used as the discriminator and generator of the network loss alternately. In the testing phase, the license plate characters are segmented and used as the input of the generator in turn, and the deblurring results are combined according to the originalorder of the license plate characters to obtain the deblurring license plate images. The model of the invention effectively restricts the edge of the license plate image, thereby improving the qualityof deblurring the license plate image and shortening the restoration time.

A method and system for estimating motion blur of an image associated with a moving object. The direction of one-dimensional motion blur may be estimated by inspecting a power spectrum associated with the image. A radon transform with respect to the image power spectrum is computed in the direction of the motion blur. A family of kernels with respect to the one-dimensional motion blur may then be defined utilizing a shutter triggering sequence associated with an image capturing device. The family of kernels may be modeled utilizing a modulation transfer function (MTF). Each modulation transfer function may be compared with the radon transform of the power spectrum associated with the image via a correlation function. The kernel with highest correlation with respect to the radon transform of the image power spectrum may be employed for de-blurring the image.

The invention discloses a single-image blind motion blur removing method based on a multi-scale residual generative adversarial network. The method comprises the following steps: acquiring a GoPRo paired data set, and connecting the GoPRo paired data set to form an image pair in a fuzzy-clear form; randomly cutting the training image into an image patch with the size of 256 * 256; taking the standardized image as model training input data; designing a convolutional neural network, and outputting a deblurred image; calculating the peak signal-to-noise ratio and structural similarity of the output information of the model and the clear image of the corresponding label, and performing loss optimization; and deblurring a picture with a motion blur scene in reality by using the optimized modelparameters to obtain a corresponding clear picture. Based on the convolutional neural network, the conditional generative adversarial network is adopted as a backbone network, and the fine-grained residual module is adopted as a main body module, so that the breakthrough of converting an image deblurring problem into an image translation problem and solving the image deblurring problem is realized, and important technical support is provided for subsequent operation of image deblurring.

The invention provides an image deblurring method based on camera motion detection. According to the image deblurring method based on the camera motion detection, motion parameters of a camera during exposure are firstly obtained through an inertial navigation device, and a rough movement locus of the camera is estimated according to the motion parameters; then a blurring core of an image during the exposure is obtained through the multi-look geometry according to the movement locus; finally, a sharp image which is deblurred is obtained according to the Wiener filtering algorithm. According to the image deblurring method based on the camera motion detection, the moving condition of the camera during the exposure can be accurately detected, the blurring core can be generated fast and accurately according to the movement locus, and finally the relatively sharp image can be obtained though inverse filtering.

The invention provides an attention mechanism-based image blind deblurring method and system, and the method comprises the steps that a multi-scale attention network directly recovers a clear image inan end-to-end mode, wherein the multi-scale attention network adopts an asymmetric encoding and decoding structure, and the encoding side of the encoding and decoding structure adopts a residual dense network block to complete feature extraction and expression of an input image by the multi-scale attention network; a plurality of attention modules are arranged on the decoding side of the encodingand decoding structure, the attention modules output preliminary restored images, and the preliminary restored images form a pyramid-type multi-scale structure of the image; the attention module further outputs an attention feature map, and the attention feature map models the relationship between the long-distance areas from the global perspective to process the blurred image; dark channel priorloss and multi-scale content loss form a loss function, and the loss function is used for reversely optimizing the network and is not self-optimized.

Provided is a method and apparatus for deblurring a non-uniform motion blur in an input image, that may restore a clearer image by dividing a large scale input image into tiles corresponding to partial areas, selecting, among the divided tiles, an optimal tile for a partial area most suitable for estimating non-uniform motion blur information, and effectively removing an artifact in an outer portion of a tile through padding of each tile.

The invention provides a method and a device for monitoring the abnormal behavior of a motor vehicle on an expressway. The method comprises the following steps of dividing and setting a monitoring area; decoding two paths of split-screen video image streams of a holographic high-definition video cameras into a single-frame image; performing deblurring and background refreshing processing on the image; detecting the motor vehicle, and tracking the motor vehicle; analyzing the driving behavior of the motor vehicle in the image; rotating a high-definition dome camera to the position of the motor vehicle to capture an image; identifying the license plate number of the motor vehicle in the image, combining the identified license plate number of the motor vehicle and the abnormal behavior of the motor vehicle into a warning statement for field playing, and simultaneously giving an alarm. According to the method and the device, the abnormal behavior of the motor vehicle on various rapid expressways is automatically identified for accurate detection and capturing of images, data and the like, and the warning statement is wirelessly transmitted and broadcast through a loudspeaker on the spot to timely prompt a driver to normally drive the motor vehicle, so that the occurrence of a fatal traffic accident can be effectively avoided, and the life and property safety of people is protected.

The invention discloses an image blind deblurring method based on edge self-adaption. To solve the problems that as for an existing total variation deblurring algorithm, edges and details of images are easily blurred, a de-mean gradient total variation canonical model is built, weighting coefficients are calculated in an iterated mode by means of local variance self-adaption of gradients of the images, and the ability of the deblurring algorithm to restore the edges and the details of the images. The image blind deblurring method comprises the following steps that (1) a blurred image is input, solutions to a gradient-region clear image and a blurring kernel are obtained alternately, and the initial blurring kernel of the blurred image is obtained; (2) the initial blurring kernel is used for conducting primary non-blind deblurring on the blurred image, and an initial clear image is obtained; (3) clustering is conducted on the initial clear image, the mean value and the weighting coefficient in the de-mean canonical model are updated, and a solution to the blurring kernel is obtained again; (4) the new blurring kernel is used for conducting secondary non-blind deblurring so as to obtain a clear image. Experimental results show that the image blind deblurring method based on edge self-adaption has better deburring effect than the prior art and can be used for image restoration.

Methods are described for efficient reconstruction of MRI data. In one practice, new reconstruction algorithms for non-uniformly sampled k-space data are presented. In the disclosed algorithms, Iterative Next-Neighbor re-Gridding (INNG) and Block INNG (BINNG), iterative procedures are performed using larger rescaled matrices than the target grid matrix In BINNG algorithm, the sampled k-space region is partitioned into several blocks and the INNG algorithm is applied to each block. In another practice, a novel partial spiral reconstruction (PFSR) uses an estimated phase map from a low-resolution image reconstructed from the central k-space data and performs iterations, similar to the iterative procedures with INNG, with an imposed phase constraint. According to yet another practice, an off-resonance correction is performed on matrices that are smaller than the full image matrix. All these methods reduce the computational costs while rendering high-quality reconstructed images.

An integrated method for both super-resolution and multi-frame demosaicing includes an image fusion followed by simultaneous deblurring and interpolation. For the case of color super-resolution, the first step involves application of recursive image fusion separately on the three different color layers. The second step is based on minimizing a maximum a posteriori (MAP) cost function. In one embodiment, the MAP cost function is composed of several terms: a data fidelity penalty term that penalizes dissimilarity between the raw data and the super-resolved estimate, a luminance penalty term that favors sharp edges in the luminance component of the image, a chrominance penalty term that favors low spatial frequency changes in the chrominance component of the image, and an orientation penalty term that favors similar edge orientations across the color channels. The method is also applicable to color super-resolution (without demosaicing), where the low-quality input images are already demosaiced. In addition, for translational motion, the method may be used in a very fast image fusion algorithm to facilitate the implementation of dynamic, multi-input / multi-output color super-resolution / demosaicing.

A multi-aperture imaging system determines depth map information. A series of image frames of a scene are captured. The frames include a normal image frame and at least one structured image frame. The multi-aperture imaging system determines edge information of an object in the scene using a deblur technique and the normal image frame. The multi-aperture imaging system determines fill depth information for the object based in part on the at least one structured image frame. The multi-aperture imaging system generates a depth map of the scene using the edge depth information and the fill depth information.

A method of using an image capture device to identify range information for objects in a scene includes providing an image capture device having an image sensor, a coded aperture having circular symmetry, and a lens; storing in a memory a set of blur parameters derived from range calibration data; capturing images of the scene having a plurality of objects; producing a set of reference edge images using the blur parameters from the stored set; providing a set of deblurred images using the captured image, the reference edges and each of the blur parameters from the stored set; and using the set of deblurred images to determine the range information for the objects in the scene.

The invention provides a video content replacement method and system based on an adversarial generative network. The method comprises the following steps: extracting a source target image in a video frame; performing semantic segmentation on the source target image; performing data enhancement on the source target image by using the image transformation operation; inputting the semantic segmentation images paired by the enhanced data set into a network training generative adversarial model; accurately detecting an ROI region of the source target image and carrying out cutting and straightening; taking the ROI image as a generation model input to obtain a generated target image and a synthetic mask; controlling edge smoothing and deblurring of the generated target image by using mask control after Gaussian blur; adjusting the hue of the generated target image to be consistent with that of the frame source target image by a histogram matching strategy; generating video frame anti-jitterprocessing; and fusing the source image and generating a target image. Compared with the prior art, the method and system has the characteristics and advantages of easiness in operation, low cost, good effect, short consumed time and the like.

Appropriate deblurring is carried out on a partially blurry image. Edge detection means detects edges in 8 directions in a reduced image, and block division means divides the reduced image into 16 blocks. Analysis means analyzes images in the blocks based on edge characteristic quantities thereof, and judges whether each of the block images is a blurry image. The analysis means obtains a width of blur, a degree of shake, and a direction of blur in each of the blurry block images as blur information, and parameter setting means sets parameters for correction based on the blur information. The parameter setting means also sets a correction strength in such a manner that the strength becomes larger as the width of blur becomes larger, according to the width of blur in the blur information.

The invention discloses a multi-scale non-local regularization blurring kernel estimation method. In image based blind de-blurring problems, posterior sharp images and blurring kernels are unknown, which is a challenging difficulty in image blind de-blurring. According to the method, on the basis of image block posteriori, a multi-scale non-local regularization strategy is introduced. The strategy is added on the two most rough layers of image pyramid calculation blurring kernels, the understanding space range is limited effectively, and accordingly, the blurring kernel calculation is evolved towards the accurate direction, clear blurring kernels are calculated, the basis is laid for the next step of sharp image calculation, image blind de-blurring is achieved through an image non blind de-blurring method, and the experiment result shows that the method is more effective and better than some existing advance methods.

The invention discloses a minimal object measurement system calibration method based on a general imaging model. A planar coordinate system is established on the plane of a calibration board, three calibration board images of different attitudes are obtained by a camera and then are subjected to deblurring, and pixel coordinates of a center of circle are obtained by ellipse fitting. The method also comprises steps that (3) a general imaging model is used to find a corresponding light equation on a world coordinate system of each pixel of the images obtained by the camera; (4) a phase position of the calibration board is obtained through an eight-step phase shifting method and a Gray code method, the coordinates of the center of circle obtained by the camera are matched with the coordinates of the center of circle of a projector according to the phase position, and a light equation of the projector, corresponding to each pixel, is obtained using the same method adopted by (3); and (5) an intersection point of light found in the camera and the projector of each pixel is computed to obtain three dimensional coordinates of an object. The invention improves the general imaging model, enables equipment to be simple and easy to implement, and gets rid of restrictions of a displacement bench in acquisition of the precise displacement.

The present invention relates to a method and system for obtaining a point spread function for deblurring image data captured by an imaging device comprising a motion sensor. First, motion path values indicating the motion of the imaging device during the exposure time are acquired. The motion path values of the imaging device are then projected onto the sensor plane and for each sensor pixel the projected motion path values are integrated over time. Said integrated value represents for each sensor pixel an initial estimate of the point spread function. Optionally, the size of the point spread function can also be estimated based on the distance of the focused object and taken into account during the projecting step.

The invention discloses a blind camera shake deblurring method based on the L0 sparse prior, and belongs to the technical field of digital image processing. The blind camera shake deblurring method is a method for deblurring blurred images caused by camera shaking, and various space-unchanged camera shaking blurred kernels, namely the point spread functions, can be estimated. The blind camera shake deblurring method solves the problem that a current variational bayes estimation method is high in computing complexity and solves the problem that a current maximum posteriori estimation method lacks strict optimization theory supports. The blind camera shake deblurring method comprises the steps of firstly, introducing remarkable edge sparse prior based on the L0 norm, and using the iterative hard threshold compressed method to achieve recessive automatic prediction of remarkable edge characteristics, secondly, introducing camera shake blurred kernel sparse prior, and using the iterative repeated weighted least square method to achieve rapid estimation of the blurred kernels, and finally, using the image non-blind deblurring method based on super-Laplacian prior to obtain a high-quality deblurred image. The flow diagram of the blind camera shake deblurring method is shown in the figure 1.

The invention discloses a simultaneous positioning and mapping method for an autonomous mobile platform in a rescue scene, which is applied to a search and rescue autonomous mobile platform and can beused for positioning and mapping in extreme environments such as a fire rescue scene, a complex dangerous accident scene and the like. The method is mainly divided into three parts of sensor information processing, pose estimation and pose correction. The method comprises the following steps: firstly, performing operations such as deblurring and feature extraction on omnibearing image informationacquired by a four-eye camera module; secondly, fusing with the measurement information of the inertial navigation sensor and estimating the pose, and finally, carrying out global pose optimization and correction. According to the method, a full-automatic end-to-end defogging method is designed based on the convolutional neural network, a post-processing step is not needed, the application rangeis wide, and the method can be applied to defogging of indoor and natural scenes at the same time. And meanwhile, a relatively stable binocular vision SLAM method based on a feature point method is adopted to fuse IMU sensor data to perform positioning and mapping of the autonomous mobile platform.