173results about How to "Eliminate ghosting" patented technology

The invention relates to the technical field of head up display, in particular to a head up display system on an automobile. The head up display system comprises a projection light source, laminated glass and a transparent nanometer film, wherein the transparent nanometer film comprises at least one laminated structure of high reflective index layers/low reflective index layers which sediment sequentially outwards from the surface of an inner glass board; the projection light source is used for generating P polarized light, the P polarized light enters the transparent nanometer film, the reflectivity of the P polarized light from the transparent nanometer film is not lower than 5%, and the incident angle of the P polarized light ranges from 42 degrees to 72 degrees. Only a reflection image of the transparent nanometer film can be observed when a reflection image on the laminated glass is observed visually, a reflection image of an uncoated surface can be neglected, and therefore the ghosting problem of the reflection image of the head up display system is avoided; by means of the head up display system, application and popularization of an HUD technology are facilitated, the safe driving environment is provided, the overall structure and process are simple, the cost is low, and the head up display system is safe, durable, attractive and elegant.

The embodiment of the invention discloses a high-precision map data processing method and a high-precision map data processing device. The method comprises the following steps of performing at least twice collection on map data along a target track so as to respectively obtain at least twice original map data; respectively recognizing objects in the original map data, and extracting object features; matching the object features in each original map datum; using an object with the matched object features as a reference substance; and performing fusion processing on each original map datum. The embodiment of the invention has the advantages that the objects in the original map data are recognized; the object features are extracted; the object features in each original map datum are further matched; the object obtained after the matching is used as the reference substance; and the reference substance is used for performing fusion processing on the original map data. The problem of offset of objects in a map image is solved; the ghost problem of the map image is solved; and the definition of the map image is improved.

The invention provides a beaming synthetic aperture radar laser imaging method, which is based on SAR imaging theory and based on the signal type difference between synthetic aperture laser radar SAL and synthetic aperture radar SAR to improve prior synthetic aperture radar SAR imaging method, to provide an imaging method suitable with synthetic aperture laser radar SAL. The inventive method comprises first processing direction pretreatment on the initial data prepared for imaging, processing Doppler frequency shift compensation on the data after pretreatment, processing point multiplication on the pretreated data and a structure phase factor exp(-j2f alpha), varying standard on the data after the Doppler frequency shift compensation, processing distance compression and direction compression on the data after the standard varying process, to obtain image data. The invention has the advantage of high image resolution, which is suitable for imaging small ground target.

The invention requests to protect a neural network nonuniformity correction method based on scene statistics, belonging to the infrared focal plane detection field. The invention provides the method aiming at the shortcoming that the traditional neural network correction method is difficult to eliminate low-frequency spatial noises. The method comprises the following steps: initializing image related matrices and parameters; detecting and compensating blind pixels according to the image pixel gray value; carrying out nonuniformity correction on image offset by a scene statistics method; judging the regional attributes of the pixels by a neural network correction method according to the standard deviation thresholds of correction errors, and carrying out nonlinear gain correction on the images which are corrected based on scene statistics and contain no low-frequency spatial noises. By the method, the correction effects of the desired image signals are good, target fade-out and ghostingare inhibited, and the changes of the background images hardly affect the correction effects. The method can be widely applied to image detection and processing.

A liquid crystal display device comprising a including signal correcting means correction for correcting a level of an original image signal to a level with which transmittance in a steady state of the pixel with the original image signal is attained within one frame period, a horizontal driving means for driver applying a voltage in correspondence with the corrected image signal to a liquid crystal material, and an illumination device for illuminating the display panel with a plurality of light emitting regions thereof, said the light emitting regions sequentially turns turning on and off in synchronization with the application of the corrected image signal, while holding a definite time delay thereto.

The invention provides a method for preprocessing high speed platform ultra-high resolution stop and reserves (SAR) data, mainly solving the problem that the traditional SAR imaging technology has low echo imaging precision of the high speed platform ultra-high resolution SAR. The realizing process comprises the following steps of: firstly, judging whether an SAR echo data azimuth frequency spectrum is overlapped, carrying out azimuth defuzzification treatment on the SAR echo with overlapped azimuth frequency spectrum; and then carrying out a go-stop movement assuming judgment on the SAR echo data, carrying out a phase compensation on the SAR echo data which do not satisfy the go-stop movement assuming judgment, and transforming the SAR echo data into the SAR echo constructed based on the go-stop movement assumption. The invention develops the application range of the traditional SAR imaging technology, through the preprocess of the invention, the high speed platform ultra-high resolution SAR data can continuously use the traditional SAR imaging technology assumed based on the go-stop movement assumption for imaging; and the invention can be applied to the ultra-high resolution SAR imaging process of the satellite-based, the missile-borne and the hypersonic unmanned planes.

A display apparatus includes a plurality of scan lines, a plurality of data lines, a plurality of pixel transistors, a plurality of pixel electrodes, a gate driver, a source driver and a discharge circuit. The data lines are intersected with the scan lines. Each of the pixel transistors is electrically coupled to a corresponding scan line and a corresponding data line, and each of the pixel electrodes is electrically coupled to a corresponding pixel transistor. The gate driver is electrically coupled to the scan lines, and the source driver is electrically coupled to the data lines. The discharge circuit is electrically coupled to the gate driver and the data lines. The discharge circuit starts when the display apparatus is turned off, to control the gate drive for turning on the pixel transistors simultaneously, and make the pixel electrodes be electrically communicated with a reference voltage.

The invention relates to a multi-exposure image fusion method, which solves the technical problems of incomplete detail information retention and incomplete ghosting elimination, and comprises the following steps of: 1, constructing an initial weight map by using exposure brightness and chroma information of a multi-exposure image sequence; 2, firstly, moving object detection is conducted on the low-dynamic image sequence, a moving area is calculated, and then a ghosting eliminating method based on super-pixel segmentation is used for eliminating ghosting; 3, completing ghosting correction onthe initial weight map in the step 1; 4, performing normalization processing on the weight map after ghosting correction in the step 3; and 5, constructing a weighted Gaussian pyramid according to theweight graph in the step 4, constructing a Laplacian pyramid of a low-dynamic image sequence, defining detail gain items, calculating a fusion pyramid, and performing image reconstruction according to the fusion pyramid to obtain a fusion image HDR. The problem is well solved, and the method can be used for image processing.

The invention relates to a transmission electron microscope specimen, which comprises a substrate on which an interest region is arranged on. A dielectric layer and a metal layer are sequentially formed on the substrate and the interest region. The difference between the contrast of the dielectric layer and the contrast of the interest region is larger than a preset value. The preparation method of the transmission electron microscope specimen comprises the steps of: providing the substrate and forming the interest region on the substrate; forming the dielectric layer on the interest region and the substrate; forming the metal layer on the dielectric layer, wherein the difference between the contrast of the dielectric layer and the contrast of the interest region is larger than the preset value; and sectioning the transmission electron microscope specimen, wherein the transmission electron microscope specimen comprises the interest region, the dielectric layer and the metal layer. The preparation method of the transmission electron microscope specimen is simple and feasible and clearer and more accurate transmission electron images can be obtained by using the transmission electron microscope specimen.

The invention belongs to the technical field of digital images, and particularly relates to a grid-optimization-based registration method suitable for free view angle image splicing. In recent years,due to the application prospect in the fields of virtual reality, medical imaging and the like, the panoramic image is greatly concerned. The image registration is mainly used for obtaining position parameters of each image in a panorama. Traditional image registration needs to execute a series of operations such as feature point extraction, matching, homography matrix solving and camera parametercorrection. According to the grid-optimization-based registration method, a grid optimization-based registration method is used for replacing a classic registration technology, so that the speed is obviously improved, and the grid-optimization-based registration method is suitable for splicing images with large parallax shot at a free visual angle. According to the grid-optimization-based registration method, ORB rapid feature extraction is used for obtaining feature points of an image, and a strategy of step-by-step coarse and fine matching is introduced, and finally, three constraint termsbased on grid optimization are introduced, and an error function minimization method is used for obtaining optimal registration parameters between the images.

The invention discloses a head-up display device. The head-up display device comprises a host machine, a display device, a curved-surface semi-transparent reflecting screen, a voice interaction unit and a driver detection device, wherein the host machine is used for transmitting an image signal to the display device, the display device is arranged at one side of the host machine and used for displaying the image signal, the curved-surface semi-transparent reflecting screen is used for receiving an image displayed by the display device and displaying the image after ghosting is removed, the image is displayed at the other side by penetrating the semi-transparent reflecting screen at one side of the host machine, the voice interaction unit is arranged in the host machine in advance, and the driver detection device is arranged in the host machine. The invention provides an intelligent vehicle-mounted device which supports whole-course voice interaction, monitors a driver in real time, gives an alarming in advance and has a good display effect, and the defects that a head-up display device is lack of voice interaction operation, is inconvenient to be placed in a vehicle and is lack of special driving emotion analysis in the prior art are overcome.

The invention discloses a video based movement target detecting method. The method comprises the following steps: obtaining an initial background through an averaging method; initiating a vibe algorithm through the initial background, and building a background model; comparing each pixel in current frame through the background model to obtain background and foreground; if the times that any pixelpoint is continuously treated as the foreground are beyond the threshold, replacing in the background model as a new background sample, and updating the background model in the previous step; and if the times that any pixel point is treated as the foreground are less than the threshold, treating as the movement target and outputting, thus realizing the detection of the movement target. According to the method, the initial background is obtained by the multi-frame averaging method and replaces the first frame of the video to build the background model, and the background sample is updated in real time in the following detecting process, so that ghost and shadow can be effectively removed; and the detection in later period can be accurately carried out.

The invention discloses a lens and an imaging device.The lens comprises a front lens set close to an object space, a reflecting element, a diaphragm and a rear lens set close to an image space.The front lens set comprises a first lens body and a second lens body.The rear lens set comprises a third lens body, a fourth lens body, a fifth lens body, a sixth lens body and a seventh lens body.The front lens set is a negative-refractive-index lens set.The rear lens set is a positive-refractive-index lens set.The reflecting element is arranged between the second lens body and the third lens body.The diaphragm is arranged between the third lens body and the fourth lens body.According to the lens and the imaging device, vignetting is avoided due to a large view field, imaging illumiance is consistent on the view field, scene light intensity energy can be absorbed to the maximum extent, meanwhile, a parasitic light forming light path is considered in the lens, ghosts generated by the lens bodies are removed by controlling the structure of the lens, and the signal to clutter ratio of a system is increased.

The present disclosure provides a circuit for discharging parasitic capacitance in a display panel with common-anode topology having a plurality of light emitters, as well as a circuit for charging parasitic capacitance in a display panel with common-cathode topology. In the common-cathode topology, the circuit includes a three-terminal device having a gate, a source, and a drain, wherein one of the source and the drain is electrically coupled to a common cathode of the light emitters, and a mechanism for controlling the three-terminal device, the mechanism being electrically coupled to the gate. Shortly after a previously selected light emitter is unselected, the mechanism turns on the three-terminal device to form a conductive path between the source and the drain. The mechanism turns off the three-terminal device after a voltage at the common cathode is increased to a predetermined voltage level or after a maximum period of time lapses.

The invention discloses a flame detection and positioning method based on multi-feature fusion and stereoscopic vision and relates to the field of computer vision. According to the method, an improvedVibe algorithm and an OHTA color splitting method are utilized to extract a suspected flame region; shape features, morphological overlapping degree features and area change rate features of a targetin the suspected flame region are extracted, so that a multi-feature fusion model is established to obtain the probability of flame existence, and whether a flame exists is detected according to theprobability; and finally flame positioning is performed according to the flame detection result, an SSDA stereoscopic matching algorithm and the principle of a parallel binocular stereoscopic vision system. The method has high anti-interference, robustness and instantaneity, the flame detection and positioning result is beneficial for positioning and extinguishment, and fire loss is lowered to themaximum extent.

The invention relates to a process for marine seismic exploration at least two different depths by means of streamers, each including a set of hydrophones, the acquisition of data at the different depths being performed in a single exploration region, and in which the position of the streamers and the hydrophones when acquiring seismic data at a first depth is independent of the position of the streamers and hydrophones when acquiring seismic data at a second depth.

The invention provides a human motion detection method suitable for a depth image. The method comprises the following steps: firstly, dividing the image into an upper layer and a lower layer, wherein the upper layer and the lower layer are used for creating a background model by different neighborhoods, and a reference model is added while the background model is created; secondly, adjusting the parameter of the difference threshold value of an image lower-layer algorithm and comparing pixels with the background model for pixel classification in a following video; thirdly, updating the background model by different updating modes based on the classified pixels; finally, denoising false detection points. According to the method, the identification and detection rates of a human body are remarkably increased.

The invention discloses a target tracking method based on Vibe and Camshift algorithms and relates to the field of computer visions. The method comprises the steps of firstly, detecting a moving target by use of an improved Vibe algorithm; then, tracking an external rectangle of an acquired target by use of Kalman filter to initialize a Camshift-tracking search window; and finally, performing target tracking by use of the Camshift, judging whether the shielding and the shielding degree exists in the target according to a target area obtained by tracking calculation and a Bhattacharyya coefficient, and performing corresponding target tracking according to the tracking shielding situation. The method is an automatic tracking method, the exact moving target is extracted, a ghost occurring inthe detection process is eliminated, the interference of a background color is eliminated for target tracking, the situation of deformation of the target can be adapted in the target tracking processand the correct tracking can be performed when the target is shielded.

The invention discloses a Vibe and disparity map background subtraction method-based moving target detection method, and relates to the field of computer vision. The method comprises the steps of firstly building a Gaussian model according to a disparity map sequence, and performing moving target detection by adopting a disparity map background subtraction method; secondly performing moving targetdetection based on an improved Vibe algorithm; and finally performing AND operation on two moving target detection results to obtain a final moving target region, continuing to update a background model, and performing moving target detection of the next frame. A monocular vision-based target detection method and a binocular vision-based target detection method are combined; the complete moving target can be extracted; the problem that the monocular vision-based moving target detection is easily influenced by lighting and shadows is improved; and the ghost phenomenon can be eliminated.

The invention discloses a transparent roller for laser imprinting. The transparent roller comprises a roller body, a roller hub and a roller shaft. The roller body comprises a quartz glass transparent roller and a plastic transparent roller which is arranged on the quartz glass transparent roller in a sleeved mode and is not in contact with the quartz glass transparent roller. The quartz glass transparent roller and the plastic transparent roller are driven in a synchronous mode through a servo motor. The transparent roller is provided with at least one mother set with an engraved laser pattern. The transparent roller further comprises a shading layer which is arranged above the plastic transparent roller. A light transmitting pattern of the shading layer is matched with the laser pattern of the mother set. Correspondingly, the invention discloses a laser imprinting device used for the transparent roller. Point-to-point location special set pattern imprinting transferring can be achieved, the fringing phenomenon of four sides of a transferred pattern is removed, imprinting transferring pattern accuracy is improved, and pattern surface quality is improved.

The invention discloses a real-time stitching method for a high-definition video. The method mainly comprises the following steps of synchronous video decoding, initialization, frame-by-frame stitching and the like. Firstly, a high-definition original video of a network camera is collected, and the video is synchronously decoded. For a first frame of video image, the initialization is carried outand includes the steps of video registration based on sub-regions, finding of an optimal stitching line set, generation of a look-up table and the like. After the initialization is completed, the frame-by-frame stitching is carried out and includes the steps of coordinate mapping of video images, detection of a moving object, stitching line update, video fusion and the like, and finally a stitchedvideo is output. In addition, a parallelization strategy is adopted for each step in video decoding and frame-by-frame stitching stages. According to the method, the real-time requirement of high-definition video stitching can be met, and a stitching frame rate is not less than 30 fps for a double-path 720P video; and meanwhile, the phenomena of ghosting, blurring, exposure inconsistence and thelike in the stitched video are avoided.

The invention relates to a secondary echo suppression and geomorphologic map fusion method based on sounding side-scan sonar, comprising two steps: (1) secondary echo suppression: a piece of sounding side-scan sonar passes through a plurality of receiving array elements to eliminate secondary sea echo using a beam-forming technology and a Chebyshev weighting method; and (2) geomorphologic map fusion: a side-scan geomorphologic map acquired by a single receiving array element and a side-scan geomorphologic map acquired by multiple receiving array elements are fused, wherein there are mainly single-array element side-scan maps near the position right below, and there are mainly multi-array element weighted side-scan maps at other positions. The problem of ghosting in conventional side-scan sonar geomorphologic maps is well eliminated. A single-array element side-scan map and a multi-array element side-scan map are fused effectively to get a higher-quality side-scan geomorphologic map, which facilitates interpretation of the submarine geomorphology.

The invention discloses an unstructured light field rendering method. obtaining a light field sampling image, taking the sampling viewpoint position as the vertex of the basic triangular mesh, the camera parameters of the sampling viewpoint and the geometric model of the to-be-rendered scene, recording the depth value of the scene under each sampling viewpoint, splitting the sampling viewpoint triangular mesh, and generating an independent sub triangular mesh for each vertex; Recording geometric information of the scene under the to-be-rendered viewpoint; And rendering the sampling viewpoint triangular mesh set, reprojecting to the viewpoint by utilizing the geometric information, sampling the light field image, and mixing the sampling results of the viewpoints. According to the method, the influence range and weight of sampling data are determined through the triangular mesh formed by the viewpoint positions when the sampling light field is split, the scene geometric model is re-projected, bilateral filtering is carried out on light field sampling based on the depth, and therefore the scene image under any viewpoint is rendered. Light field sampling data of various sources can befully utilized, and a vivid rendering result can be generated in real time.