174results about How to "Improve image processing efficiency" patented technology

The invention belongs to the technical field of security and protection equipment manufacture, and particularly relates to an intelligent security and protection system for a smart new town. An accesscontrol system is a 4G face capturing camera deployed on the inspection spot position of a key area for realizing ultrahigh-density face capture; under a condition that the system is not in direct contact with equipment, the camera collects face dynamic pictures at fixed frequency, and a face recognition algorithm is applied to each frame of detected face image; an intelligent face recognition algorithm module is embedded into the 4G face capturing camera to realize face recognition on the front end of the camera; the 4G face capturing camera with the built-in face recognition algorithm module nearby processes and collects data on front-end camera equipment on an edge side; and near-end service is provided on one side of a data source, an application is initiated on the edge side, a quicknetwork service response can be generated, a time delay and transmission cost are reduced, and the transmission cost of a broadband and a load stored on the China Unicom cloud can be lowered.

The invention discloses an unmanned aerial vehicle collected image-based construction fugitive dust pollution source automatic monitoring and influence scope prediction system and method. The system comprises an aerial photographing system and a ground supervising and processing system, wherein the aerial photographing system comprises an aerial photography rotor aircraft, a mobile device and a control terminal; the mobile device and the control terminal communicated with the aircraft in a wireless manner, the aircraft is controlled to collect pollution source monitoring zone images, and the ground supervising and processing system comprises an image processing unit and a user terminal; the image processing unit is used for preliminary screening, distortion correction and smoothing of qualified images among aerial photos; the image processing unit is also used for pollution source RGB color characteristic enhancement, edge detection, pollution source existence determination and recheck, suspected pollution source characteristic extraction, pollution source RGB color characteristic value extraction, suspected pollution source characteristic comparison, image splicing, and pollution source distribution analysis and influence scope prediction; the user terminal can display complete pollution source monitoring area images, pollution source distribution map and influence scope prognostic maps.

The invention relates to a distance measuring method of a significant target in a binocular image. The distance measuring method of the significant target in the binocular image aims to solve the problem that an existing target distance measuring method is low in the processing speed. The method includes the steps that step 1, significant feature extraction is conducted on the binocular image through a visual significance model, and a seed point and a background point are marked; step 2, a weighed graph is established for the binocular image; step 3, a significant target in the binocular image is partitioned through a random walk image partitioning algorithm by means of the seed point and the background point in step 1 and the weighed graph in step 2; step 4, key point matching is conducted on the significant target separately through an SIFT algorithm; step 5, the significant target distance is worked out by applying a parallax matrix K' worked out in step 4 into a binocular distance measuring model. The distance measuring method can be applied to distance measurement of the significant target of the image in front of the vision in the intelligent automobile running process.

A medical image processing system including a medical image generating apparatus for generating medical image data of a radiographed subject; a suspicious region candidate detecting apparatus for detecting a suspicious region candidate from the generated medical image data, the detecting apparatus being connected with the generating apparatus through a communication network; a network monitoring section for monitoring a communication load of the communication network; and a time setting section for setting an image transmission time for transmitting the medical image data from the generating apparatus to the detecting apparatus on the basis of a monitoring result of the communication load. The generating apparatus transmits the medical image data to the detecting apparatus at the set image transmission time.

The invention provides a fluorescence in situ hybridization (FISH) image parallel processing and analysis method. The method employs parallel processing, detection of chromosome fluorescent labeling points and edge detection and segmentation of nucleuses can be performed simultaneously, and the processing time is reduced; and a watershed algorithm based on adaptive shape labels is employed so that the segmentation precision of the adhered nucleuses in FISH images is substantially improved, the detection accuracy of relative positions of the nucleuses in tumor cells and the chromosome fluorescent label points is improved, and online real-time detection of the tumor cells can be realized.

The invention discloses a semi-automatic digital method of a non-standard map. The method comprises the following steps of: 1) scanning to obtain a grid image, and performing geographical space registration and correction and binarization; 2) establishing a linear vector layer and assigning a geographical space reference system consistent with the corrected grid image; 3) overlapping the linear vector layer on the grid image after the binarization, identifying and capturing the center lines of the lines in the grid image, automatically tracking the captured center lines of the lines in the grid image one by one, and performing line breakage and endpoint artificial judgment on each automatically tracked line, and performing vectorization to obtain a vector image; 4) performing attribute assignment, check and correction; and 5) re-establishing the topological relation of the vector image, and checking and correcting the topological errors such as overlapping and gap to obtain a vector image after digitalization. The method disclosed by the invention has the advantages of good image processing accuracy, high image processing efficiency, low processing workload, short processing period and wide application range.

The invention discloses a method and device for video storage. One specific execution mode of the method comprises the steps: obtaining two adjacent picture frames of a video from a first storage region for video buffering; comparing pixel points in preset regions of the two adjacent picture frames and judging whether there are changes of the pixel points or not; enabling the later picture frame to be stored in a second storage region if there are change of the pixel points; comparing pixel points in non-preset regions of the two adjacent picture frames and judging whether there are changes of the pixel points or not if there are no change of the pixel points; enabling the later picture frame to be stored in the second storage region and enabling the pixel points with change to be combined in a preset region if there are change of the pixel points. The execution mode can improve the efficiency of image processing.

The present invention provides an image processing apparatus which can verify whether or not an image has been able to be correctly read and easily correct an image which has not been able to be correctly read. The image processing apparatus has a reading section 105, an image processing section 110 for image-processing a read image, and a CPU 110 for controlling the entire apparatus. The image processing section 110 has a detection section 101 for detecting an image in a different state, a verification section 102 for displaying an image in a different state, and a correction section 106 for correcting the image. The image processing section 110 has a preview function for displaying the read image on a display section 108, and a correction function for correcting the image. The image processing apparatus can correct only a tilted image without rereading all images.

The present invention discloses an image processing system, a method thereof and a device to solve the problem of poor size compatibility of a processed image in an ISP architecture in the prior art. The system comprises an image sensor for obtaining an image, a sampling module for sampling a received video image according to the image resolution configured by an image sensor and the maximum line resolution processed by an ISP module, the ISP module used for receiving a video image transmitted by the sampling module, carrying out image processing on the video image, reading the static state image currently stored in a memory according to the maximum line resolution processed by an ISP module, carrying out image processing on the static state image, and sending the processed image to the memory, and the memory for storing the image. In addition, the configuration of processing the size of the image by the ISP module is not needed, the need of line buffer in the image processing process is reduced, and the problem of the poor size compatibility of the image processed by the ISP module is solved.

The invention discloses a multi-view human facial image gender identification method and a device, wherein the method comprises the following steps: a step of matching by classifiers and a step of identifying human facial image gender. The step of matching by the classifiers comprises the following steps: S11, obtaining a plurality of sample human facial images from a human facial image library; S12, extracting sample characteristics of all sample human facial images and fuzzy matching the first classifier according to the sample characteristics; and S13, testing the sample characteristics by the first classifier and screening a target characteristic, and precisely matching the second classifier according to the target characteristic. The step of identifying human facial image gender comprises the following steps: S21, collecting video images of the target characteristic containing human faces to be detected; S22, extracting the target characteristic of the human faces to be detected; and S23, processing the target characteristic by the second classifier and identifying the human facial gender. The method and device disclosed by the invention can improve the identification rate of human facial images collected from different views.

The invention discloses a digital high-fidelity image collecting and processing method of stone cave temple wall paintings. The method is characterized in that through photography collection, image treatment and image storage, by utilizing a cavalier perspective and parallel translation principle and adopting an autonomously designed photography collecting system and a unique cross-shaped skeleton image splicing processing method, the data is stored safely, normalized stone cave cultural relic wall painting digitalization is carried out, thousands and even more original images photographed for the same wall surface can be spliced and fused into a traceless and perfect high-fidelity full-scale integral imaging picture, and the purposes that the precious cultural heritage can be stored permanently and used sustainably for human beings are achieved.

The invention discloses a video analysis based recognition method and device. The recognition device comprises a video preprocessing module, an identity library module, a tracking and detecting module and an alarm information module, and is characterized in that the video preprocessing module is used for acquiring a video data packet, converting the data packet into an image and performing binarization processing on the image; the identity library module is used for extracting personnel identity attribute information and personnel key part feature information, storing the personnel identity attribute information and the personnel key part feature information in a separated manner and forming a personnel identity library; the tracking and detecting module is used for tracking and extracting feature target information of personnel key parts on the binary image, comparing the extracted feature target information with personnel key part features in the personnel identity library and acquiring a judgment result; and the alarm information module is used for storing and pushing the judgment result in a mode of setting data types. The recognition device solves a problem of recognition and warning for strangers, the image processing efficiency is high, a lot of storage space is saved, the cost is low, a personnel management and control solution is provided for a video monitoring service in special areas, and early warning for abnormal events and convenient query after the events are achieved.

The invention relates to an underwater image color correction method based on a GAN network. The method comprises the steps of training a Cycle GAN network to complete the style conversion from an image in the air to an underwater image; generating the corresponding underwater images from the images in the air by using the trained Cycle GAN network, wherein the images in the air and the corresponding underwater images form a paired data set; constructing a UGAN network; using the paired data set and the actually shot underwater image as a verification set to train the UGAN network; carrying out color correction on the underwater image by using the water removal model, fusing the UV channel of the input underwater image and the Y channel of the output image in the air, and converting the fused UV channel of the input underwater image and the Y channel of the output image in the air into an RGB image to be output. According to the present invention, the data set, matched with the underwater image corresponding to the image in the air, of the generated image in the air is real and available, the image quality can be improved after the underwater image is dewatered, the processing speed is fast, and the efficiency is high.

The present invention provides a sheet processing apparatus comprising a sheet holding device capable of storing supplied sheets, a first sheet stacking tray on which sheets are stacked at a downstream side of the sheet holding device in a sheet conveying direction and on which the sheets are subjected to processing, a sheet discharging device for discharging the sheets stacked on the first sheet stacking tray, a second sheet stacking tray on which the sheets discharged by the sheet discharging device are stacked, and a controller for controlling the sheet discharging device and the sheet holding device if sheet jam occurs at an upstream side of the sheet holding device, in such a manner that, after the sheets on the first sheet stacking tray are discharged onto the second sheet stacking tray, the sheets stored in the sheet holding device are discharged onto the first sheet stacking tray.

The embodiment of the invention discloses an image processing method and apparatus, and a storage medium. The method comprises the steps of obtaining a video file of a single-channel video stream, each frame of image in the video file being spliced by a colored image and a gray image corresponding to the colored image; performing decoding on the each frame of image in the video file to generate adata image corresponding to the each frame of image; performing segmentation on the data image to generate a colored image and a gray image corresponding to the each frame of image; performing synthesis on the colored image and the gray image to obtain a synthetic image corresponding to the each frame of image; and generating a target video file according to the synthetic image, and displaying thetarget video file. The effect that the processing only needs to be performed on the video file of the single-channel video stream is implemented, operations such as synchronous hardware decoding on atwo-channel video, copying an image from a GPU (Graphics Processing Unit) to a CPU (Central Processing Unit) to synthesize and the like are avoided, and the efficiency of a terminal to perform imageprocessing is improved.

The invention aims to provide a method and device for providing a photographed image. The method comprises the steps that a providing device acquires an initial preview image in a photographing preview state; according to image feature information of the initial preview image, one or more pieces of additional information corresponding to the initial preview image is determined; according to the initial preview image and in combination with the additional information, an image previewing device corresponding to the initial preview image is generated; according to the preview image, the photographed image corresponding to the preview image is provided for a photographing user and / or a device of a target user. Compared with the prior art, the method and device have the advantages that one-key operation of a photographing device or application is achieved, the photographing-first and editing-second mode is changed, the operating steps of the user are decreased, the image processing efficiency is improved, and user experience is improved.

The invention discloses an image processing method and an image processing device based on artificial intelligence, equipment and a storage medium, which belong to the technical field of computers. According to the image processing method of the invention, image distribution difference between sample human tissue images in two fields is determined, the human tissue image in the source field is subjected to image conversion, the target sample image obtained after conversion conforms to an image distribution rule of the human tissue image in the target field and carries the corresponding targetprocessing results, and manual re-annotation is not needed, so that the manual annotation step is omitted, the labor cost is reduced, the model training efficiency is improved, and the image processing efficiency is further improved; and the human tissue image of the target field is acquired and a training sample can be automatically obtained through carrying out image conversion on the human issue image, and the image processing method is more universal and higher in applicability.

An embodiment of the invention provides an image bilateral filtering method and device. The method comprises the following steps: after obtaining pixel characteristic parameters of a pixel to be filtered in an image and neighborhood pixels thereof, carrying out table look-up operation based on the pixel characteristic parameters to obtain relevant parameters, and calculating pixel contribution sum and pixel weighting sum of all neighborhood pixels; and finally, through the pixel contribution sum, the pixel weighting sum and the pixel characteristic parameters of the pixel to be filtered, calculating a bilateral filtering result of the pixel to be filtered, and carrying out image processing on the pixel to be filtered according to the bilateral filtering result. The method is characterized by converting index calculation into table look-up operation, looking up a preset reciprocal table to obtain reciprocal of the pixel weighting sum and converting division operation into multiplication operation, thereby preventing a lot of floating-point precision parameter division and index operation, improving operation speed of the bilateral filtering result, reducing image processing waiting time and improving image processing efficiency and user experience.

The invention discloses a fast previewing system of a surveillance video. The fast previewing system comprises a dynamic frame extracting module for extracting a dynamic frame image containing a moving target from the surveillance videos, a dynamic frame filter module for filtering repeated dynamic frames, a dynamic frame play module for playing the filtered dynamic frame images, a dynamic frame searching module for searching dynamic frame images in a gallery, a dynamic frame jump module for jumping the dynamic frame images in the galley to the original video and starting playing the original video on a dynamic frame appearing part, and a dynamic frame storage module for storing the filtered dynamic frame images into a database. The fast previewing system disclosed by the invention has the beneficial effect that the dynamic frame image containing the moving target can be quickly previewed in the video before a user replays the video, and the video can be replayed from the part when the interested moving target is seen; static frames can be filtered and attention is concentrated on a significant moving target, so that the video viewing efficiency is improved.

A sheet processing apparatus includes: a buffer unit which stores plural supplied sheets with upstream edges in a conveying direction thereof aligned; a processing tray on which sheets discharged from the buffer unit are stacked; and an oscillation roller pair and a return roller which convey the sheet stacked on the processing tray to bring the sheet into abutment against a stopper for receiving the upstream edge of the sheet. The buffer unit is adapted to align the upstream edges of only sheets to be stored before a sheet to be supplied last among the sheets to be stored.

An image processing method and apparatus are provided which are applied to scheduling nodes in an image processing system. The image processing system also includes multiple image processing nodes. The method comprises: when an image processing task is received, selecting one image processing node from the plurality of image processing nodes as a target image processing node according to a presetimage processing scheduling strategy; assigning the image processing task to the target image processing nodes to so that the target image processing nodes perform the image processing task. The technical scheme of the application can improve the image processing efficiency.

The invention provides a defect detection method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a to-be-recognized image and a depth image of a to-be-detected object, the pixel positions occupied by the to-be-detected object in the to-be-recognized image and the depth image being the same; inputting the to-be-recognized image into a defect detection model to obtain a defect area output by the defect detection model; determining depth information corresponding to pixel points in the defect area according to the depth image; and obtaining a defect detection result according to the depth information. According to the embodiment of the invention, image acquisition is carried out on the to-be-detected object, defect identification is carried out on the image, after the defect area is determined, the depth information is obtained by using the depth map corresponding to the defect area, and finally whether the defect is the target defect is determined according to the depth information. According to the embodiment of the invention, the defect detection is realized by processing the image, and the target defect is determined in combination with the depth information, so that the defect detection efficiency is improved.

A fish character measurement device comprises an image collection device and an image processing system. The image processing system comprises: an image reading module configured to read images; a feature identification module configured to identify the contour and feature points of a fish body in the images; an image identification module configured to generate a measurement template drawing having the same length and width with the fish body image pixels, wherein the measurement template drawing comprises a measurement rectangle externally connected with the fish body contour to measure at least one measurement vertical line being parallel to the width direction of the fish body in the rectangle, the measurement rectangle is regulated to allow the measurement rectangle to perform overlapping display on the fish body contour in the images, and the at least one measurement vertical line is arranged on the at least one feature points; and a dimension calculation module configured to calculate a distance between any two of the vertical lines in the measurement rectangle to take the distance as the feature dimension of the fish body according to the coordinate values of the four corners of the measurement rectangle and the coordinate values of the intersection point of the vertical lines in the measurement rectangle and one side of the measurement rectangle. The measurement efficiency and the measurement accuracy of the fish body characters can be improved.

The invention provides an image batch-processing method, a device thereof and electronic equipment. The image batch-processing method, the device thereof and the electronic equipment relate to the field of image processing technology. The method is executed by means of a neural network. The image batch-processing method comprises the steps of acquiring a plurality of to-be-processed first images;performing splicing processing on the plurality of first images, thereby forming a second image; and processing a second image for obtaining a processing result. The image batch-processing method, thedevice thereof and the electronic equipment can effectively improve image processing efficiency.

The invention provides a rendering method for a virtual grass cluster in a game. The method comprises the following steps: determining stress points of grass leaves of the virtual grass cluster in a current game scene; when at least one game skill is rendered in the current game scene, determining force field information of the game skill; generating a scene force field graph matched with the current scene according to a preset four-channel color buffer and the force field information of at least one skill; and rendering the virtual grass according to the scene force field graph and the stresspoints. Therefore, a multi-skill coverage rendering effect is achieved, a vivid grass cluster effect is rendered, grass clusters are prevented from deforming in the rendering process, and the image processing efficiency and the user experience are improved.

The invention discloses an image processing method and system based on machine vision, and relates to the technical field of image processing. The method comprises the following steps: acquiring an RGB image; carrying out gray processing on the obtained RGB image; removing the background and extracting an R-B difference grayscale image as a target image; processing pixel points in the target imageto obtain an image detection area: extracting a reference image of the target image by using a deep neural network in machine vision, and processing the sub-target image and the reference image in the detection area to obtain a binary image of the sub-target image and the reference image; and calculating a binary image of the reference image and a binary image of the sub-target image by utilizinga deep neural network in machine vision to carry out difference operation, obtaining a defect value, and carrying out image defect judgment. By means of the technology, the image processing efficiency of machine vision is effectively improved, and the time of research and development personnel is saved.

The invention discloses an image processing algorithm for an AOI (automated optical inspection)-based bullet appearance defect detection system. The image processing algorithm comprises the following steps: (1) obtaining an original image of a bullet appearance by an AOI system; (2) processing the original image by a median filtering method, removing noise, and smoothening an original image signal; (3) stretching the image by linear grey level transformation; and (4) segmenting the image by Otsu to obtain a binarization image. The original image obtained by the AOI system is subjected to processing, such as smooth denoising, image enhancement, image segmentation and the like by an algorithm with a high operation speed, so that the original image is converted into the binarization image, bullet appearance defects can be more obviously highlighted and can be more conveniently extracted and identified in a subsequent process, and image processing efficiency is greatly increased.

A contact image-sensing module having fingerprint-scanning function is described. The image-sensing module has a flex / rigid composite substrate having a first rigid circuit substrate, a second rigid circuit substrate, a third rigid circuit substrate, and at least one flex circuit board. The first rigid circuit substrate, the second rigid circuit substrate, and the third rigid circuit substrate respectively and electrically connect to the flex circuit board in order to form a first composite substrate, a second composite substrate, and a third composite substrate. Each of the second composite substrate and the third composite substrate respectively is not located in the same plane and has a predetermined angle with respect to the plane where the first composite plane is located.

The invention relates to an SSE2 (streaming SIMD extensions 2nd) instruction set based image interpolation method. The method includes: selecting a current pixel point, and loading image data in a neighboring domain of the current pixel point into an xmm register; interpolating a component G of each pixel point of a row GR and a row BG in the neighboring domain of the current pixel point; outputting data of the component G after interpolation; interpolating a component B and a component R of each pixel point of the row GR in the neighboring domain of the current pixel point on the basis of the interpolated component G data; interpolating a component B and a component R of each pixel point of the row BG in the neighboring domain of the current pixel point on the basis of the interpolated component G data; outputting interpolated BGR-format 24-bit image data. Outputted Bayer-format Raw data are acquired aiming at cameras, operations of multiple pixel points in an instruction execution period are realized on the basis of the SSE2 instruction set technology, and consequently calculation complexity is lowered, integral processing efficiency is improved, and more time is reserved for other post-processing procedures aiming at images.

The invention discloses a gridding visualization method for scanning laser radar detection data. The gridding visualization method specifically comprises the steps of: 1, defining a region according to a detection scanning distance and a detection angle of scanning laser radar detection data; 2, performing coordinate conversion on polar coordinates of the scanning laser radar detection data, and converting the polar coordinates into a Cartesian product coordinate form V (m, n); 3, standardizing the data V (m, n) by adopting a normalization method to obtain U (m, n); 4, performing interpolationfilling on a data missing position in the U (m, n) by utilizing a vertical and horizontal interpolation method; 5, creating and initializing two-dimensional space color matrix data C (m, n) corresponding to visual graph sizes m and n, and mapping data G (m, n) into the C (m, n) representing numerical value intensity; and 6, drawing numerical values of the C (m, n) to pixel positions by adopting aprogram visualization method by utilizing a corresponding relationship between pixels and the C (m, n). The gridding visualization method improves the precision of data interpolation.