53results about How to "Reduce the number of images" patented technology

The invention discloses a high dynamic-range synthetic imaging method and device, aiming at providing a method for synthesizing two pieces of images with lower dynamic ranges into a piece of image with higher dynamic range. The high dynamic-range synthetic imaging method comprises the steps of: shooting a same scene at different exposure time to obtain two pieces of different exposed images; respectively converting the two different exposed images to be in exposure value spaces according to response curves of a shooting device; weighted-synthesizing scene brightness which is calculated by utilizing exposure value and exposure time according to brightness of the two pieces of images; and simulated-exposing the synthesized scene brightness through the response curves to obtain the image with high dynamic range. According to the invention, two pieces of different exposed images are synthesized, and thus the difficulty of image alignment is reduced; moreover, the operand and the required storage space in the processing procedure are reduced, so that the high dynamic-range synthetic imaging method and device can be applied to a small-sized handheld device; and at the same time, the image with high dynamic range is rich in details at light parts and shadow parts, and the color and the details of the image are in conformity with those of the original scene.

The invention discloses a rapid image splicing method based on wide-angle lenses. The method includes the steps that firstly, distortion correction is conducted on images acquired through the wide-angle lenses; secondly, cylindrical surface mapping is conducted on the images on which distortion correction is conducted, and a two-dimensional discrete coordinate mapping table is generated for each wide-angle lens; thirdly, the two-dimensional discrete coordinate mapping table of each wide-angle lens is loaded, and distortion correction and cylindrical surface mapping are sequentially conducted on the distortion images acquired through the wide-angle lenses according to the tables; fourthly, translation parameters between adjacent visual angle images on which cylindrical surface mapping is conducted are acquired, and the adjacent visual angle images are registered; fifthly, the registered adjacent visual angle images are fused, wherein Laplacian pyramids are established for registered adjacent visual angle image sequences respectively, linear fusion is conducted on a high-frequency pyramid image overlaying region and a low-frequency pyramid image overlaying region, and then a fused image of the adjacent visual angle images is obtained. The method has the advantages of being high in splicing speed and real-time performance.

The invention discloses an image based method for identifying railway overhead-contact system bolt support identifying information. According to the method, a digital image processing method is used for automatically identifying bolt support numbers and kilometer posts in railway overhead-contact system touring images. The method comprises the steps of 1) conducting coarse positioning on bolt supports; 2) conducting fine positioning on bolt supports; 3) conducting bolt support number detection and digital dividing; 4) conducting kilometer post detection and digital dividing; and 5) conducting offline training and online identifying on bolt support numbers and kilometer post numbers. The method can efficiently and automatically identify bolt support numbers and kilometer posts of a railway overhead-contact system, is applicable to routing inspection of overhead-contact systems of high-speed railways and conventional railways, can effectively help inspectors to quickly position specific bolt supports, can reduce the number of images which are required to be interpreted manually and can improve the working efficiency.

The invention discloses an automatic segmentation system for a digestive tract internal image. The system comprises an image preprocessing module, a feature extraction module and an image segmentation module, wherein the signal output end of the image preprocessing module is connected with the signal input end of the feature extraction module, and the signal output end of the feature extraction module is connected with the signal input end of the image segmentation module. Through the automatic segmentation system, the time for a doctor to search for images of different parts of the digestive tract can be shortened, the workload is reduced, and work efficiency is improved.

The invention discloses an on-machine vision detecting method of a rotating cutting tool and a detecting device thereof. The method comprises the following steps that a first camera is used for capturingan image of the front cutting face of the rotary cutting tool; the image is subjected to processing and analyzing to obtain position angles of cutting edges; a second camera is controlled to rotateto the lateral cutting face according to the angels and the axis of the second camera is vertical to the lateral cutting face, and then shooting is conducted; images of the lateral cutting faces of the cutting edges are obtained; the images of the front cutting face and the lateral cutting faces are subjected to vision analyzing to obtain the abrasion condition of the cutting tool. The detectingdevice comprises the first camera and the second camera perpendicular to each other and mounted at a bracket, and further comprises a rotating motor driving the bracket to rotate. A rotating shaft ofthe rotating motor is concentric with the axis of the first camera. On-machine detection can be achieved. Abrasion areas of the lateral faces of the cutting tool can be accurately shot, and detectingefficiency is improved.

Embodiments of the application disclose a defect detection method, system, and apparatus. The method can be applied to the field of cloud computing. Specifically, a specific embodiment of the method includes: separately inputting images captured under multiple photographing conditions to a target detection model, to obtain multiple first detection results output by the target detection model; foreach of the images captured under the multiple photographing conditions, determining, based on a first detection result output by the target detection model for the image, whether a target object presented in the image has a defect, and if it is determined that there is the defect, using the image as a suspected defect image; and inputting the suspected defect image to a pre-trained neural networkmodel for detection, to obtain a second detection result. The embodiments of the application can filter out the suspected defect image through a fast unsupervised model, thereby reducing a quantity of images input to the neural network model and improving efficiency of defect detection.

The embodiment of the invention provides a license plate recognition method, device and system. The license plate recognition method comprises a step of detecting a collected license plate image by using a preset image segmentation algorithm and determining the position of each character in the license plate image, a step of determining a tilt angle range of the license plate image according to the position of each character in the license plate image, a step of performing rotation correction on the license plate image according to the tilt angle range to obtain a corrected license plate image, a step of acquiring vertex coordinates of the corrected license plate image and transforming the vertex coordinates through perspective transformation to obtain a converted license plate image withthe same height of the character, and a step of identifying each character in the converted license plate image to obtain a recognized license plate. Through the scheme, the calculation efficiency andsuccess rate of license plate recognition can be improved.

The invention relates to an on-line product characteristic detection method and system, in particular to a visual detection method and device based on a virtual reality technology. The invention provides the visual detection method and device by aiming at problems in the prior art. The system is characterized in that a circular ring identification coordinate system is firstly obtained through transformation according to an image coordinate system; after an image of an object to be detected is shot, a digital model of a standard object is virtually projected to a previous circular ring coordinate system, and a virtual standard object image is generated; and then, a comparison detection method is used for judging a difference between a real shot image and the virtual standard object image so as to determine whether a product is a qualified product or not. The virtual standard object image generated each time can be directly stored for subsequent detection and comparison and also can be dynamically generated. In addition, in order to reduce an amount of the virtual standard object images, features which can be used for estimating the spatial position of an object image to be detected can be firstly extracted from the object image to be detected, then, only the virtual standard object images on a plurality of possible positions can be generated, a data processing amount is effectively lowered, and detection time is effectively shortened.

The invention relates to the field of image fusion, and provides a panoramic image splicing method, a terminal device and a computer readable storage medium. The method comprises the following steps:performing quality screening on an acquired image sequence so as to save the image sequence satisfying the quality requirements in the image sequence; filtering out key frames from the image sequencemeeting the quality requirements according to a feature matching algorithm to obtain an optimal image set; and splicing together the images in the optimal image set to obtain a panoramic image. According to the technical scheme provided by the invention, on one hand, the panoramic image finally obtained through splicing is high in image sharpness and high in quality; and on the other hand, the amount of images needed for panoramic image splicing is reduced, so that the processing efficiency in the panoramic image splicing is raised.

The invention discloses a method and a device for online preview of a three-dimensional model. The method comprises the steps of locally pre-rendering a picture in multiple view angles for the three-dimensional model; and responding to the online preview of the three-dimensional model. According to the method and the device for the online preview of the three-dimensional model, real data of the model do not need to be loaded locally, a user is prevented from contacting the real data and the data safety is protected; an optimization scheme for rendering of the view angles is provided, thus thenumber of images needing to be drawn is greatly reduced, the data transmission size is decreased and the data transmission time is shortened, the real-time rendering of a three-dimensional scene is realized, and the usage experience of the user is improved; the display performance requirement for mobile terminal equipment is reduced; and the limit of a browser is avoided so that the preview of a large scene becomes possible.

The invention provides a hardware architecture of a medical endoscope artificial intelligence system and an image processing method. A data processing module is arranged to screen out images with theconfidence exceeding a predetermined threshold value and transmit the images to a remote server, and therefore the communication load is reduced; and the data size needing to be remotely transmitted by an endoscope containing the data processing module is correspondingly decreased, and then the communication load of the endoscope is reduced. The data processing module is arranged to mark predetermined feature areas to display specific image information to related staff, the image recognition workload is reduced, and the working efficiency is improved; and the endoscope containing the data processing module also can display the images with the specific image information to the related staff, and the working efficiency and the detection real-time performance are improved. The data processingmodule is arranged to screen out the images with the confidence exceeding the predetermined threshold value, only the part of images with the confidence exceeding the predetermined threshold value need to be stored later, and therefore the needed storage space can be effectively decreased.

The invention relates to a point cloud model construction method and device, equipment and a storage medium. The method comprises the steps of obtaining a first feature of a panoramic image in a panoramic image set; determining at least one group of image pairs in the panoramic image set and a corresponding matching result according to the first features, the image pairs comprising two panoramic images matched with the first features, and the matching result indicating a corresponding relation between the first features of the two panoramic images; and constructing a point cloud model according to the at least one group of image pairs and the corresponding matching result. The panoramic images are adopted for matching, and the point cloud model is further constructed according to the matching result, so that the number of the images of the image set can be reduced, the matching efficiency and the modeling efficiency are improved, the range of the space corresponding to the panoramic images is relatively large, the matching effect between the panoramic images can be improved, and the precision and the quality of constructing the point cloud model are improved.

The invention provides an organ image constructing method and system based on a voxel model. The method and system are used for solving the technical problems that in the prior art, the CT cost is high, image reading is difficult, and information cannot be screened in advance. The organ image constructing method based on the voxel model comprises the steps of S1 model rendering and S2 model rasterization. The organ image constructing system based on the voxel model comprises an original model module, an image processing module and a display module; by implementing the technical scheme, more information is displayed by images, the number of images needed by diagnosis is reduced, meanwhile, direct display can be achieved through a displayer, and the cost is greatly lowered; a doctor can seethe perspective drawing of an organ from multiple angles, understanding is easy, and the diagnosis efficiency is improved; the displayed voxel type can be selected, the system can supply various imagetypes, and the diagnosis accuracy is improved.

The invention discloses a high dynamic-range synthetic imaging method and device, aiming at providing a method for synthesizing two pieces of images with lower dynamic ranges into a piece of image with higher dynamic range. The high dynamic-range synthetic imaging method comprises the steps of: shooting a same scene at different exposure time to obtain two pieces of different exposed images; respectively converting the two different exposed images to be in exposure value spaces according to response curves of a shooting device; weighted-synthesizing scene brightness which is calculated by utilizing exposure value and exposure time according to brightness of the two pieces of images; and simulated-exposing the synthesized scene brightness through the response curves to obtain the image with high dynamic range. According to the invention, two pieces of different exposed images are synthesized, and thus the difficulty of image alignment is reduced; moreover, the operand and the required storage space in the processing procedure are reduced, so that the high dynamic-range synthetic imaging method and device can be applied to a small-sized handheld device; and at the same time, the image with high dynamic range is rich in details at light parts and shadow parts, and the color and the details of the image are in conformity with those of the original scene.

The invention discloses a pathological section training image set acquisition method, which comprises the following steps: acquiring a target pathological section, acquiring at least one visual fieldimage of the target pathological section, each visual field image comprising at least one currently labeled cell; acquiring cell labeling information labeled by at least one currently labeled cell inthe at least one view image; obtaining a to-be-screened training image set according to the at least one view image and the cell labeling information labeled by the at least one currently labeled cellin the at least one view image; and acquiring an image evaluation standard, and performing image evaluation on each view image in the to-be-screened training image set according to the image evaluation standard so as to screen and obtain the training image set meeting the image evaluation standard. According to the method, a large amount of invalid content contained in the image training set is removed, memory occupation of the image training set can be effectively reduced, and the method is very suitable for obtaining the data source of the artificial intelligence model. In addition, the invention also provides a training image set acquisition system, equipment and a medium.

Embodiments of the present invention provide a license plate recognition method, device, and system, wherein the license plate recognition method includes: detecting the collected license plate image through a preset image segmentation algorithm, and determining the position of each character in the license plate image; According to the position of each character in the license plate image, determine the inclination angle range of the license plate image; according to the inclination angle range, perform rotation correction on the license plate image to obtain a corrected license plate image; obtain the corrected license plate image The vertex coordinates are transformed by perspective transformation to obtain a transformed license plate image with the same height as the characters; each character in the transformed license plate image is recognized to obtain the recognized license plate. Through this solution, the calculation efficiency and success rate of license plate recognition can be improved.

The invention discloses an image interpolation method and device based on RGBD images and a multi-camera system. The method comprises the steps that camera calibration is conducted on each camera in the multi-camera system; determining an interpolation position of a new camera according to the position information of each camera in the multi-camera system, and calculating a camera pose of the newcamera according to camera calibration data; according to the projection relationship of the cameras and the pose information of each camera, calculating a plurality of initial interpolation images having a one-to-one correspondence relationship with specified images acquired by each camera in the multi-camera system; performing image fusion on each initial interpolation image to obtain a fused interpolation image; and performing pixel completion on the fused interpolation image to finally obtain an interpolation image associated with the new camera. According to the invention, the problem that a multi-view-angle video is easy to have an unsmooth impression during switching of viewing view angles due to the fact that the number of the acquisition cameras is too small is solved.

The invention discloses a panoramic depth image acquisition method and device, a panoramic depth image synthesis method and device, equipment and a storage medium, and relates to the technical field of optical imaging. The acquisition method comprises the following steps: acquiring the characteristic that a focal power value changes along with time; acquiring a field depth range corresponding to the focal power value; obtaining a target focal power value; acquiring an acquisition moment for image acquisition; and collecting a target image. The image synthesis method comprises the step of performing depth-of-field fusion processing on an acquired image to obtain a full depth-of-field image. The device comprises a time characteristic acquisition module, a field depth range acquisition module, a target focal power value acquisition module, an acquisition time acquisition module and an image acquisition module. Computer program instructions are stored on the storage medium. According to the panoramic depth image acquisition method, synthesis method, device and equipment and the storage medium, the technical problems of complex image acquisition process, large image processing data volume and low processing speed of the existing image processing technology based on the variable-focus lens unit can be solved.

Embodiments of the present invention disclose an image display method and apparatus, and electronic equipment. The method comprises: determining garbage images in the electronic equipment according to a preset garbage image determination rule; modifying filename extensions of the garbage images to filename extensions corresponding to non-image files; and scanning and displaying images in the electronic equipment. By virtue of the method, apparatus, and electronic equipment disclosed by the embodiments of the present invention, since the filename extensions of the garbage images are modified, thus when scanning the images, the garbage images are no longer scanned, so that the garbage images are no longer displayed; and since there is no garbage image in the displayed images, the number of displayed images is reduced, thereby improving image searching efficiency.

The embodiment of the invention discloses a method and device for obtaining vehicle conditions. According to the historical vehicle condition image of the vehicle, the latest vehicle condition image of the vehicle can be obtained, the latest vehicle condition image is sent to the user terminal so that the current vehicle condition image of the vehicle uploaded by the user terminal according to thelatest vehicle condition image can be obtained, and the vehicle condition information represented by the current vehicle condition image is different from the vehicle condition information represented by the latest vehicle condition image. That is to say, when the user finds that the vehicle condition of the vehicle is different from the vehicle condition represented by the latest vehicle condition image, the current vehicle condition image can be uploaded through the user terminal, and when the vehicle condition of the vehicle is the same as the vehicle condition represented by the latest vehicle condition image, the current vehicle condition image is not uploaded, so that the number of uploaded images can be reduced, the vehicle use experience of a user is improved, the workload of image analysis is reduced, and the vehicle condition obtaining efficiency is improved.

An embodiment of the invention discloses an image processing method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a target area in a picture shot by a first camera; acquiring a plurality of depths of field of a plurality of non-overlapping regions in the shot picture; acquiring m images of a shot picture through a first camera, selecting n target images from the m images according to the plurality of depth of field, and synthesizing the n target images to obtain a first high dynamic range image. Thus the image quantity of the synthesized image can be reduced on the premise of ensuring the HDR image processing effect, the system power consumption is reduced, the frame time delay is reduced, the requirement of a user for the HDRimage processing effect is met, and the shooting experience of the user is improved.

The invention relates to an image matting effect display method and device, a storage medium and electronic equipment. The method comprises the steps that under the condition that an image matting effect display instruction for a target video is obtained, the image matting effect of a first image is displayed, and the first image is a first image matted in response to the target image matting instruction in the target video; under the condition that the second image does not exist, displaying an image matting effect of the third image; wherein the second image is an image which is located behind the first image in the target video and is subjected to image matting in response to the target image matting instruction, and the third image is an image which is located behind the first image in the target video and is not adjacent to the first image. According to the embodiment of the invention, the image matting effect can be displayed at intervals of some image frames, and compared with frame-by-frame display of the image matting effect, the video image matting full view can be quickly displayed.

The embodiment of the present application discloses an image processing method, device, electronic equipment, and storage medium. The method includes the following steps: by determining the target area in the shooting picture of the first camera; acquiring multiple non-overlapping areas in the shooting picture multiple depths of field; acquire m images of the shooting picture through the first camera, select n target images from the m images according to multiple depths of field, and synthesize the n target images to obtain the first high dynamic range image, so, On the premise of ensuring the HDR image processing effect, the number of composite images can be reduced, thereby reducing system power consumption and frame delay, thereby meeting the user's demand for HDR image processing effect and improving the user's shooting experience.

The invention discloses an image based method for identifying railway overhead-contact system bolt support identifying information. According to the method, a digital image processing method is used for automatically identifying bolt support numbers and kilometer posts in railway overhead-contact system touring images. The method comprises the steps of 1) conducting coarse positioning on bolt supports; 2) conducting fine positioning on bolt supports; 3) conducting bolt support number detection and digital dividing; 4) conducting kilometer post detection and digital dividing; and 5) conducting offline training and online identifying on bolt support numbers and kilometer post numbers. The method can efficiently and automatically identify bolt support numbers and kilometer posts of a railway overhead-contact system, is applicable to routing inspection of overhead-contact systems of high-speed railways and conventional railways, can effectively help inspectors to quickly position specific bolt supports, can reduce the number of images which are required to be interpreted manually and can improve the working efficiency.

The invention relates to a key frame mark-based sequence CT image pelvic contour extraction method, which includes the following steps: CT image preprocessing; obtaining a region of interest mask; convolving the mask with an initial image to obtain an intermediate image; Arrange in spatial order, select the pixel difference between two adjacent frames of CT images as the feature distance, and use the two adjacent frames with the feature distance less than the specified threshold as similar frames. Preliminarily screen the CT sequence to obtain the candidate key frame sequence; The key frame sequence is carefully screened; the key frame bone contours are interactively marked; the GVF model is used to extract the bone contours of each layer of CT images.

The invention discloses a feature progressive finger vein image accurate matching method based on a ULBP (Uniform Local Binary Pattern) and SURF. The method comprises the following steps: S1, preprocessing a finger vein image; S2, carrying out image enhancement on the preprocessed image; S3, matching the extracted features with feature information in a database, and outputting a matching result. A coarse-to-fine finger vein recognition algorithm is realized; the problems that a single ULBP algorithm is small in gray scale range and low in recognition precision, and time consumption is too long when matching is conducted through SURF feature points are solved.

The invention discloses a structured light coding method based on wrapped phase forward and reverse. The method comprises the following steps: 1) constructing a structured light measurement system; 2) adjusting the positions of a camera and a projector, and storing calibration parameters; 3) generating a random positive and negative sequence to be used; 4) synthesizing three phase shift projection images containing positive and negative sequence information according to the generated sequence; 5) outputting the three phase shift projection images to the surface of the measured object, and collecting three stripe grating images after the surface of the measured object is modulated and deformed; 6) carrying out phase unwrapping on the three stripe grating images to obtain an initial wrapping phase; 7) adjusting the initial wrapping phase to obtain an adjusted wrapping phase and a period number; 8) calculating a full-field phase, and carrying out three-dimensional reconstruction; the method is suitable for most traditional phase shift profilometry, and the number of images needed by projection is greatly reduced under the condition that the precision, advantages and experiment requirements of an original method are reserved.

The invention discloses a speckle pattern-based mirror surface object three-dimensional measurement method, which comprises the following steps of firstly, generating a speckle pattern by using a computer, respectively projecting the speckle pattern on a reference mirror surface and a measured mirror surface by using an LCD screen, respectively acquiring by using an imaging system, and calculating the displacement of speckles by using an open-source 2D DIC application program Ncorr, acquiring the height of the measured mirror surface by directly establishing the relation between the speckle displacement and the height, and recovering the three-dimensional shape. Compared with a traditional sine fringe measurement method, the method has the advantages that patterns needing to be shot are greatly reduced, and the measurement efficiency is improved; by establishing the direct relation between the speckle displacement and the height, the gradient integration process is omitted, error accumulation in the gradient integration process is avoided, and the defect that a traditional method cannot measure discontinuous mirror objects is overcome; the method is simple in calculation, flexible in control and simple and convenient to operate, has high robustness in measurement of mirror objects, and has potential application prospect and practical value in precision measurement.

The invention discloses a precise positioning method for a target reporting system, which comprises: overlapping a thermal imaging sensor and an image sensor for framing; using the thermal imaging sensor to continuously acquire thermal images of target paper, and analyzing the thermal images to obtain initial coordinate information P of the impact point and time information T; use a visible light image sensor to continuously acquire image H, select several frames before T and after T in image H, a group of image information I centered on P; analyze I to obtain precise coordinate information K of the impact point; Extract the target paper image feature information in the image H; obtain the target report information according to the precise coordinate information K and the target paper image feature information. Further, a target reporting system is also disclosed. The invention can effectively eliminate false alarms generated when other non-bullets penetrate the target paper, and the analysis difficulty and required time are greatly reduced, the implementation is simpler, the use is more convenient, and the acquisition of target information is more accurate and reliable.