1121 results about "Median filter" patented technology

Defects such as dirt, dust, scratches, blemishes, pits, or defective elements or pixels in a CCD, scanner, photocopier, or image acquiring device are dynamically detected by processing a plurality of images via a computer. A pristine object of calibration is not required. Stationary components of the video images are found and detected so as to produce a low false alarm probability. Text segmentation and measurement of total deviation based on variability related to high-frequency components of the video image are employed to prevent applying the process or method to printed text or graphics. Additional techniques optionally employed are median filtering, sample area detection, and dynamic adjustment of scores. In special cases, only moderately blank documents are used. The dynamic defect detection allows defect compensation, defect correction, and alerting the operator of defects.

One or more median filter circuits are used to filter radio link control measurements corresponding to one or more radio link parameters of interest, such as received signal quality or round trip delay measurements, such as might be used by a base station to trigger mobile station handoff. As such measurements are particularly susceptible to measurement outliers arising from rapid but short-lived changes in radio link propagation paths, for example, the application of median filtering to such measurements is particularly advantageous. That is, by operation of median filtering, which is a non-linear filtering process, outliers in a stream of control measurements, such as are caused by instantaneous changes in channel fading or other propagation phenomena, are discarded rather than averaged in with the other measurements. Non-linear filtering as implemented by exemplary median filtering does not impair or otherwise limit the bandwidth of the underlying control measurements.

In a method for suppressing interferences while detecting objects in a target area, a transmitter transmits a sequence of pulses into the target area, and a receiver detects the resulting reflection signal of the pulses reflected from the objects, within successive time windows that are referenced to the moment of transmitting an individual pulse and thus represent distance gates. The time spacing between the successive individual pulses is variable and randomized according to the pseudo-noise principle within predetermined limits, and the time windows are adapted accordingly. The received reflection signal may be sampled, digitized, digitally pre-processed and digitally filtered in the individual distance gates. A non-linear digital filter, preferably a sliding median filter, is used for the filtering to suppress transient disturbances. The median is determined from an odd number of consecutive sampled values of a reflection signal detected within a distance gate.

A method is provided for coherently tracking a radio signal including a plurality of digitally modulated reference subcarriers. The method comprises the steps of receiving symbols transmitted on the reference subcarriers, combining the reference subcarrier symbols with a known reference sequence conjugate to produce a plurality of samples, median filtering the samples to produce filtered samples, and smoothing the samples for each of the reference subcarriers over the plurality of reference subcarriers to produce a coherent reference signal estimate for each of the subcarriers. A receiver for coherently tracking a radio signal including at least one digitally modulated reference carrier is also provided.

Methods and systems for hierarchically layered adaptive median motion vector smoothing are disclosed. Aspects of one method may include generating motion vectors (MVs) for video pictures at each level of a hierarchical motion estimation process, where each level may use different resolution video pictures. The MVs may be smooth filtered at each level to reduce or remove spurious MVs. The smooth filtering may comprise scalar median filtering and / or vector median filtering. The smooth filtering may receive as inputs a plurality of MVs, for example, the MV being filtered and eight MVs from the surrounding video blocks. The vector costs of the nine MVs may be compared to a threshold vector cost, and those MVs with costs above the threshold value may be discarded. The threshold value may be dynamically changed based on, for example, content of the video pictures and / or to adjust a window size of the smoothing filter.

The invention disclose a bar code image identification method, belonging to the field of computer information automatic identification technology, the method comprises the following steps: S1. carrying out gradation to a colored image which is captured by an Android mobile phone camera, wherein the colored image is a bar code image; S2. conducting median filtering processing to a grayed image; S3. conducting binaryzation processing to the image which carries out the median filtering processing; S4. conducting bar code positioning to the image which carried out the binaryzation processing; S5. conducting geometric distortion correction to the positioned image; and S6. decoding the corrected image, so as to obtain a code word of the image. In the invention, the precision of the bar code identification is improved and the complexity is reduced.

The invention discloses an automatic fruit identification method of an apple picking robot on the basis of a support vector machine. An apple orchard colored image under natural illumination is collected, and a vector median filter is adopted to pretreat the apple colored image; after pretreatment, the image is cut with the method of combining region growing with color characteristics; the color characteristics and the geometry characteristics of the cut apple colored image are respectively extracted; apple fruits are indentified with the mode identification method of the support vector machine; and finally, the fruits are accurately positioned. The identification method of the support vector machine of the invention which integrates color characteristics and shape characteristics has higher apple fruit identification precision rate than the precision rate when only color characteristics or shape characteristics are adopted and has better identification effect; the algorithm is easy to realize, the operation time is short, and the identification performance is superior to a commonly used neural network method, so that the identification method shows advantages in small sample learning.

The invention relates to a device and method for detecting defects on the surface of a solar cell. The device comprises a structural member and a detection circuit system. The structural member is composed of a display frame, an observation platform and an installation frame. The detection circuit system comprises an illumination unit, an image collection unit, an image processing unit and a display unit. In the detection process, an unfilled corner defect image is obtained by means of a low-angle annular white LED light source and by conducting image segmentation, wavelet transform and two-dimensional 7*7 pixel field median filtering on an image; for a crack, a defect image of the crack is obtained by conducting two-dimensional median filtering, wavelet transform, image binaryzation, edge detection and morphology operator processing on the image, the detection recognition result is obtained, and the image processing unit finally transmits the processing result to an upper computer of the display unit. The method and the device have the advantages that compared with a manual visual detection and infrared scanning detection method, detection efficiency and detection accuracy are greatly improved, the method is easy to operate and practicable, a large amount of labor force is saved, and the labor intensity is lowered.

The invention discloses a method and device for acquiring a Kinect depth image. The method comprises the steps of sampling static scenes continuously by means of the Kinect to obtain a plurality of frames of depth images and colored images; carrying out median filtering and noise reduction on the depth images; carrying out edge detection on the first frame of colored image and the depth images treated with median filtering by means of the Canny operator; carrying out primary segmentation on the depth images by means of a user index value of Kinect depth data to obtain a mark sheet, wherein pixels with the same mark value belong to one region in the mark sheet; carrying out primary repairing on the depth images by means of a region growing method with edge limitation considered and updating the mark sheet according to edge information of the depth images and the colored images and a depth image segmentation result; carrying out image enhancement on the depth images by means of trilateral filtering according to the updated mark sheet. According to the method and device for acquiring the Kinect depth image, edge inconformity of the depth images and the colored images is effectively overcome, cavities in an original Kinect depth image are repaired, noise is eliminated at the same time, and the depth image which is good in quality can be obtained finally.

An image sensing apparatus comprising: a CCD, optical system for forming an image on the CCD; A / D converter for converting an image signal outputted by the CCD to a digital signal; color interpolator for performing color interpolation on the digital signal converted by the A / D converter and generating image data on R, G and B color planes; color space converter for converting RGB color space to YUV colorimetric system; and median filter for reducing pseudo color components, generated by the color interpolator, by controlling color difference signals U and V. By cutting high frequency components of the signal, filtered by the median filter, then thinning out YUV signals and performing JPEG compression, pseudo color components generated by color interpolation can be reduced.

The invention belongs to the technical field of two-dimensional code identification, and provides a rapid positioning and identifying method based on two-dimensional code decoding. The method comprises the following steps: performing median filter processing on an entire image; scanning the image, judging whether or not a two-dimensional code exists in the image, and if so, acquiring a two-dimensional code regional coordinate; intercepting a two-dimensional code image; performing adaptive binarization processing; performing connected area disposal on black pixel points; screening four boundaries; converting the coordinate system of the image into the coordinate system of the two-dimensional code; projecting the two-dimensional code under a new coordinate system to obtain the directivity, real boundary length and coordinate of the two-dimensional code and the barycentric coordinates of black-white blocks; judging the gray values of the black-white blocks to obtain a two-dimensional code information matrix; and acquiring decoding information. The method can adapt to various working conditions, and has the advantages of high contrast, high interference resistance and low investment cost. Automatic positioning and identifying functions can be finished by only one set of industrial camera and light source; high working efficiency is achieved; and the production requirement of an assembly line is met to the maximum extent.

The invention discloses an improved image edge detection method. The improved image edge detection method comprise steps of S1, performing smoothing processing on an image and using an improved media filter to suppress noise, S2, obtaining a difference between a horizontal direction and a perpendicular direction through first-order-partial derivatives of directions of x,y,45 degrees and 135 degrees and thus obtaining a gradient amplitude and a gradient direction, S3, performing non-maximum value inhibition on the gradient amplitude, S4, using a gradient histogram to solve a high threshold and a low threshold and then using a double-threshold algorithm to perform edge detection on the image, and S5, performing sharpening processing and connecting the edges to obtain a final edge image. The improved image edge detection method use weighted media filtering to replace gauss filtering, uses the partial derivatives of four directions, uses the gradient histogram to determine the high threshold and the low threshold, reduces detection errors, improves detection accuracy, and makes the counter of the edge image more clear and continuity better.

Conventionally, in image processing using a smoothing filter, the large amount of calculation for 25× the number of pixels takes much processing time. A computer 21 as the nucleus of image processing determines whether or not a pixel of interest is a color blur pixel, based on the intensity of a low-density element color, at steps S130 and S140. If the pixel is a color blur pixel, the computer 21 determines whether or not the color blur pixel is an edge pixel at step S150. If the pixel is not an edge pixel, the computer 21 uses a smoothing filter at step S152, while if the pixel is an edge pixel, the computer 21 uses a median filter at step S154, in image processing to reduce the color blur. This reduces the amount of calculation and realizes high-speed image processing.

The invention relates to a method for detecting surface defects of a dustproof cover of a bearing based on machine vision, which comprises the following steps: a, obtaining a gray level image of a to-be-detected bearing; b, carrying out separation on the gray level image so as to obtain a ring image of the dustproof cover of the bearing; c, carrying out grey level transformation and adaptive median filtering preprocessing on the ring image; d, carrying out threshold segmentation on the ring image by using a maximum interclass variance method, extracting an edge of the ring image of the dustproof cover of the bearing after threshold segmentation by using a Roberts edge detection operator; e, calculating the numbers of which the numerical values are 1 in the image area every two degrees for the ring image of the dustproof cover of the bearing; f, providing a template bearing image, and obtaining a deflection angle sigma through calculating; g, carrying out separation on the image so as to obtain character areas and non-character areas of the ring image of the dustproof cover of the bearing; and h, carrying out surface defect judgment on corresponding character areas and non-character areas in the ring image of the dustproof cover of the bearing according to the number of connected areas and the defect area. By using the method disclosed by the invention, automatic detection can be achieved, the visual detection workload of artificial detection is reduced, the detection efficiency is improved, and the method is safe and reliable.

A content-dependent scan rate converter with adaptive noise reduction that provides a highly integrated, implementation efficient de-interlacer. By identifying and using redundant information from the image (motion values and edge directions), this scan rate converter is able to perform the tasks of film-mode detection, motion-adaptive scan rate conversion, and content-dependent video noise reduction. Adaptive video noise reduction is incorporated in the process where temporal noise reduction is performed on the still parts of the image, thus preserving high detail spatial information, and data-adaptive spatial noise reduction is performed on the moving parts of the image. A low-pass filter is used in flat fields to smooth out Gaussian noise and a direction-dependent median filter is used in the presence of impulsive noise or an edge. Therefore, the selected spatial filter is optimized for the particular pixel that is being processed to maintain crisp edges.

The invention discloses a full convolutional neural network-based large-range remote sensing image semantic segmentation method. The method comprises three stages of data tagging, model training and result prediction; the method performs preprocessing operations of waveband synthesis, image fusion and image segmentation on remote sensing images for characteristics of high precision, large range, multispectral information and the like of the remote sensing images; the richness of samples is improved by applying a data enhancement technique to a training set; and if a data set with a label can be obtained, the data set can be used for training a model firstly, and target data model training is initialized to reduce the workload of manual labeling. For improving the accuracy of a result, themethod performs overlapping grid division on the images, performs prediction, splices prediction result images in sequence, and performs median filtering to reduce noises and unsmooth parts in the images; and finally, relatively high accuracy is achieved.

The invention relates to the field of medical image processing and mode recognition, in particular to a tubercle bacillus target recognizing and counting algorithm based on diverse characteristics. The algorithm comprises the following steps of: image preprocessing: carrying out image reinforcement and constructing median filter and Gaussian filter on a tubercle bacillus microimage; color image partition: carrying out fixed threshold partition based on HSV (Hue-Saturation-Value) color space on a preprocessed image and then carrying out adaptive threshold partition which is based on CIE L*a*b* color space and keeps a geometric shape of a target; communication block morphological analysis and target recognition: carrying out communication block analysis on the partitioned image; and tubercle bacillus target counting: estimating the quantity of tubercle bacillus targets in the image by utilizing a histogram statistics and multistrategy calculation method. The invention can effectively extract the bacillus targets in the tubercle bacillus microimage subjected to acid-fast stain from background and impurities and carry out accurate counting, thereby realizing the automation and the intellectualization of the detection of tubercle bacilli.

The invention discloses a quick motion estimation method of multi-view video coding, comprising the following steps of: carrying out spatial median filtering on a motion vector field of coded frames of adjacent views; calculating an overall parallax vector between the current view and the adjacent view by using a parallax vector field of the coded frame of the current view; selecting a reference motion vector from the filtered motion vector field by using the overall parallax vector; selecting a searching center of motion estimation by using the reference motion vector and the spatial motion vector of adjacent estimated blocks; selecting the searching range of the motion estimation by using the deviation degree of the searching center and the reference motion vector; and carrying out the final motion searching in the searching range, selecting the final motion vector of the current block, and finishing the selection of the final motion vectors of all blocks by using the same method and finishing the motion estimation of the multi-view video coding. The method is suitable for the motion estimation of the multi-view video encoding and can effectively reduce the calculation amount of the motion estimation and synchronously maintain the coding rate distortion performance.

The invention is a method and apparatus for reducing noise in an image. The method and apparatus involves calculating a plurality of directional operators, comparing the directional operators to a predetermined threshold, and applying a filter responsive to the comparing. The method and apparatus computes the directional operators by taking a weighted sum of the absolute differences between a target pixel and its surrounding pixels. The comparison signals to the method or apparatus the existence of a line or edge. If the method or apparatus detects no edge or line, the method applies a smoothing or averaging filter. If the method or apparatus detects an edge or line, the method applies a median filter in the direction with a minimum directional difference.

The invention discloses a moving body posture estimating method which can accurately position body framework points and effectively obtain the features with more expressive power in a three-dimensional moving body and is more effective and simpler in structure. The method includes the steps of firstly, preprocessing depth image data through a median filter method, and conducting part calibrating on body pixels through a Dijkstra algorithm based on geodesic distance; secondly, determining that the number of clusters of each type is three on the basis of a regional feature point extracting algorithm of a K-mean value clustering algorithm, and extracting 32 posture features to express different body postures; thirdly, obtaining position calibrating information of the framework points through PoserPro2012 software in the training stage, synthesizing 300 frames to simulate posture features of a human body, calibrating standard framework points, and calculating a linear regression model of the posture features and the framework points through the posture feature points and the standard framework points of a training sample so that the mapping relation between the posture features and the standard framework points can be obtained.

The invention discloses a saliency-based synthetic aperture radar (SAR) image airfield runway edge detection method, which comprises the following implementation steps: (1) reading in image data; (2) carrying out median filtering on the image data which is read in; (3) carrying out enhanced processing on a filtered and de-noised image; (4) obtaining a salient map of the processed image; (5) rejecting a noisy background area on the basis of the saliency information of the salient map; (6) obtaining a seed growth point to carry out region growing; (7) carrying out aftertreatment; and (8) carrying out edge positioning. According to the saliency-based SAR image airfield runway edge detection method, the airfield runway edge can be quickly detected, the airfield runway edge information can be favorably kept, meanwhile, the interference of the noisy background area is avoided, and the saliency-based SAR image airfield runway edge detection method has better instantaneity and higher detection precision and can be applied to the SAR image airfield runway edge detection.

The invention discloses an up-conversion device and method of the video frame rate. The up-conversion device comprises an input-output module, a motion estimation module, a motion vector median filtering module, a reconfiguration module, a deblocking filtering module, a DDR and controller module, a state machine controlling module and the like. The device can improve the video frame rate and generates video with high quality. The up-conversion method includes the following steps of respectively conducting motion estimation on a forward frame and a backward frame of a reconfiguration frame; conducting a multi-frame extrapolation method and a direct interpolation method or conducting a method of motion estimation of the size and self-adaptive threshold judgment of a variable block according to comparison between SAD value (the sum of absolute values of differences) and threshold of a present block; obtaining initial motion vectors through the motion estimation and updating the threshold of the present image block; filtering motion vectors which is estimated mistakenly through a median filtering method based on a time domain and a space domain; and conducting reconfiguration and deblocking filtering and outputting results.

The disclosed picture removing-noise method based on multi-template mixed filter comprises: dividing the objective picture into MXN blocks without superposition to contain every pixel in some block; defining one group of filters including the mean filter and mid-value filter, every with different removing-noise strength; defining a numerical sequence with number standing for picture information size; obtaining the noise variance of the picture, calculating the uniformity level for the picture block to thereby select one filter for finishing the removing-noise. This invention simplifies calculation, and reduces real running time.

Methods for coherent block-based motion estimation for motion-compensated frame rate conversion of decoded video sequences are provided. In some of the disclosed methods, motion vectors are estimated for each block in a decoded frame in both raster scan order and reverse raster scan order using prediction vectors from selected spatially and temporally neighboring blocks. Further, in some of the disclosed methods, a spatial coherence constraint that detects and removes motion vector crossings is applied to the motion vectors estimated for each block in a frame to reduce halo artifacts in the up-converted video sequence. In addition, in some of the disclosed methods, post processing is performed on estimated motion vectors to improve the coherence of the motion vectors. This post-processing includes application of vector median filters to the estimated motion vectors for a frame and / or application of a sub-block motion refinement to increase the density of the motion field.

Embodiments of the invention allow for cloning texture from one image location to another. The two locations may belong to the same image, or to two different images. Embodiments of the invention use a median filter to filter a source image and a target image. The filtered source image and the original source image are combined to extract texture information. Texture information is then combined with the filtered target image to generate a destination image.Embodiments of the invention may be utilized to retouch photographs, in order to correct, improve or change localized flaws or problems in images. An example is the removal of wrinkles and blemishes on a person's face. Embodiments of the invention allow for cloning the texture from a smooth skin area into a region of wrinkles, while preserving the colors and shadings of the target image.

The invention discloses a cockpit instrument automatic recognizing and detecting method, which comprises the following steps that instrument images are read; the images are sampled; nonlinear vector median filters are adopted for carrying out noise reduction processing on the images; overall and local threshold value methods are combined, instrument images are subjected to binaryzation, and binaryzation images are obtained; the images are refined, pointers are accurately detected, and the pointers subjected to refining processing become single pixel width pointers. The cockpit instrument automatic recognizing and detecting method has the advantages that an improved crossed visual model is utilized, and instrument edges are extracted; according to priori knowledge, the study training is carried out, similar features are found, and instruments are subjected to classified comparison; a gradient method is utilized, and the angle of the pointers is calculated; through the angle, the numerical value is calculated by combing the priori knowledge, and in addition, the storage display is carried out. Cockpit instruments are completely and automatically recognized and detected, the manual intervention is not needed, manpower resources can be greatly reduced, errors caused by subjective factors are avoided, and the cockpit instrument automatic recognizing and detecting method with good performance is provided.