412results about How to "Guaranteed resolution" patented technology

The invention discloses a brain tumor segmentation network and segmentation method based on a U-Net network. The tail of a contraction path of the segmentation network is connected with a spatial pyramid pooling structure; hole convolution of different scales is introduced into a network jump connection part of the segmentation network; an Add operation and original input are adopted to form a residual block with hole convolution; a receptive field of shallow feature information in the contraction path is expanded; fusing with an expansion path of a corresponding stage is carried out. The segmentation method comprises the following steps: cutting and preprocessing a training data set, then constructing a brain tumor segmentation network DCU-Net based on a U-Net network, then inputting a preprocessed two-dimensional image into a segmentation model for feature learning and optimization, obtaining an optimal parameter model of the segmentation model, and finally inputting a to-be-segmented test data set image into the segmentation model for tumor region segmentation. According to the method, the problems of over-segmentation and under-segmentation in brain tumor segmentation can be effectively solved, and the brain tumor segmentation precision is improved.

The invention belongs to the technical field of computer applying and the subfield of the machine learning theories and application and focuses on the traffic sign recognizing problem in the intelligent traffic technology. A traffic sign recognizing method based on multi-resolution convolution neural networks is characterized by solving the problem that the speed is low when the convolution neural network is used for recognizing traffic signs, two-dimensional images with different resolutions are used as input, the two convolution neural networks with the same structure are operated in parallel to carry out feature mapping and extracting, and accurate classifying and recognizing are carried out based on a weight threshold value trained by the networks. The two CNNs with different resolution branches are used for replacing a basic CNN structure, the overall and outline features can be mapped through the high-resolution image input, the local and detailed features can be mapped through the low-resolution images, the recognizing resolution is guaranteed, and the model training speed is increased.

The invention discloses a no-grid co-prime array DOA (direction of arrival) estimation method based on virtual array interpolation, and mainly solves a problem of information loss caused by the heterogeneity of a virtual array in the prior art. The method comprises the implementation steps: constructing a co-prime array at a receiving end; receiving an incident signal through the co-prime array, and carrying out the modeling; calculating an equivalent virtual signal corresponding to the receiving signal of the co-prime array; constructing an interpolation virtual array, and carrying out the modeling; constructing a multi-sampling snapshot signal of the interpolation virtual array and a sampling covariance matrix; constructing a projection matrix, and defining projection calculation correlated with the projection matrix; designing an optimization problem based on the minimization of the nuclear norm of the signal covariance matrix of the interpolation virtual array, and solving the optimization problem; and carrying out the estimation of the DOA according to the reconstructed covariance matrix of the interpolation virtual array. The method improves the DOA estimation freedom degree and resolution, and can be used for the passive positioning and target detection.

The invention discloses a co-prime array wave arrival direction estimation method based on interpolation virtual array signal atom norm minimization. The objective of the invention is to solve the problem of an information loss caused by non-uniformity of a virtual array in the prior art. The method comprises following steps that a receiving end constructs a co-prime array; based on the co-prime array, an incidence signal is received and modeling is performed; an equivalence virtual signal corresponding to the a co-prime array receiving signal is calculated; an interpolation virtual array is constructed and modeling is performed; a multi-sampling snapshot signal of the interpolation virtual array and a sampling covariance matrix thereof are constructed; a projection matrix is constructed and projection calculation related to the projection matrix is defined; an optimization problem based on the interpolation virtual array signal atom norm minimization is designed and solved; and direction-of-arrival estimation is performed according to the reconstructed interpolation virtual array covariance matrix. According to the invention, the freedom degree and the resolution of the direction-of-arrival estimation are improved and the method can be used for passive positioning and target detection.

The invention provides a method for comprehensively recovering tungsten and fluorine from minerals, namely a mixed acid of phosphoric acid and sulfuric acid is adopted for decomposing complex calcium-containing minerals containing fluorite, scheelite, apatite, and calcite, wherein the fluorite is decomposed to fluorine hydride or silicon tetrafluoride to escape, and absorption treatment is performed for preparing hydrofluoric acid or a fluoride salt; and the scheelite is transformed to phosphotungstic acid to enter into a solution, and filtrate after filtration is supplemented into the consumed sulfuric acid and the phosphoric acid after extraction of the tungsten and returned to the new-round mineral leaching. The method disclosed by the invention has the advantages of comprehensively recovering the fluorine and the tungsten from the minerals, reducing the requirements on the fluorite or the tungsten ore raw material, reducing the pressure on a mineral dressing link, improving the comprehensive recovery rate and simultaneously ensuring the decomposition rate of the fluorite and the scheelite, wherein the decomposition rate of the fluorite is above 98%, and WO3 contained in decomposition slag is reduced to below 0.5%; furthermore, a leaching agent can be recycled, so that leaching cost and wastewater emission are greatly reduced; and the method also has the advantages of simple leaching equipment, convenience in operation and easiness in realization of industrialization.

The invention provides a super-resolution imaging system and method with stable images. The method comprises the following steps: firstly, placing an imaging optical element group, an oscillating mirror, a spectroscope, an imaging lens group and an imaging camera sequentially in the imaging optical path direction of a propagation path of an observed aiming beam; connecting an input end of a super-resolution image synthesis unit, which is used for synthesizing a plurality of undersampled images transmitted by a receiving and imaging camera into super-resolution images, with an output end of the imaging camera; placing the imaging optical element group, the oscillating mirror, the spectroscope, a detecting lens group and a detecting camera sequentially in a detecting optical path direction of the propagation path of the observed aiming beam; connecting an output end of the detecting camera with an input end of a computing control unit; connecting an input end of an oscillating mirror driver with an output end of the computing control unit; connecting an output end of the oscillating mirror driver, which is used for amplifying the voltage analog quantity sent by the computing control unit into the driving voltage analog quantity of the oscillating mirror and outputting the driving voltage analog quantity to the oscillating mirror, with a driving end of the oscillating mirror.

The invention discloses an SAR (synthetic aperture radar) direction ambiguity suppression method based on antenna main lobe dominance intensity constraint and relates to synthetic aperture radar image techniques. The method includes: determining an antenna main lobe energy dominance region according to distribution conditions of SAR direction ambiguity energy within a PRF (pulse repetition frequency) corresponding to an antenna directivity diagram; completing data imaging of the antenna main lobe energy dominance region; taking the amplitude relation between the imaging result and an original full-resolution image as a judgment condition to subject the original full-resolution SAR image to selective intensity constraint to realize direction ambiguity suppression of the original image. The method overcomes the defect that traditional suppression methods are limited by scenes and target characteristics or only suitable for local image processing, and direction ambiguity suppression of large-area SAR images can be quickly completed while indexes such as SAR image resolution, peak sidelobe ratio and integral sidelobe ratio are not worsened and original phase information of the SAR images is reserved.

The invention discloses a low-cost dual-band optical zoom lens, which sequentially comprises a front lens group with negative focal power, a diaphragm and a rear lens group with positive focal power from an object side to an image side, wherein the front lens group comprises a convex-concave first lens with negative focal power, a double-concave second lens with negative focal power and a convex-concave third lens with positive focal power, the rear lens group comprises a double-convex fourth lens with positive focal power, a double-convex fifth lens with positive focal power, a double-concave sixth lens with negative focal power and a double-convex seventh lens with positive focal power, and the fifth lens and the sixth lens are glued assemblies. A variation of a 2.8-10mm focal range is realized through changing the air space between the front lens group and the rear lens group. The lens disclosed by the invention only uses seven lenses, materials of the lenses do not contain expensive lanthanide series glass and ultralow chromatic dispersion glass, has good processing properties, and effectively controls the cost. The lens is designed for a 1/2.7-inch sensor, the resolution achieves mega pixels within the overall focal range, and day and light confocal imaging is realized.

The embodiment of the application provides a laser marking method. The method comprises following steps: a control mainboard receives coordinates and grey values of target image pixel points; the control mainboard calculates the output power of the target image pixel points according to the grey values; and the control mainbaord controls a laser to output laser light aiming at the target image pixel points with corresponding output power according to the displacement of a vibration mirror. According to the application, an independent pixel point can embody the change of the grey value by the adjustment of the output power of the target image pixel points (marking points), so that the grey value of the image can be presented. The coordinates of the pixel points of the grey-scale image is calculated according to preset resolutions, and the coordinates of the pixel points are not adjusted during a marking process, so that the resolution of the grey-scale image can be guaranteed. A manner that the gray scale is embodied through the density of the marking points is abandoned, and each marking point can embody the grey scale, so that the marking efficiency is relatively improved.

The invention discloses a display panel and a display device, belongs to the technical field of display, and is used for preventing the deformation of a high light transmission area caused by the external pressure, thereby ensuring the light transmission effect of the high light transmission area and ensuring the resolution of a photosensitive assembly arranged below the high light transmission area. The display panel comprises dividing a non-display area into a first area with smaller touch signal line density and a second area with larger touch signal line density according to the wiring density of the touch signal lines in the non-display area, wherein the non-display area is provided with a first sub-support column and a second sub-support column; in a first area, a supporting component is arranged on one side, facing a second substrate, of a first substrate, orthographic projections of at least part of the first sub-support column on the first substrate are overlapped with orthographic projections of the supporting component on the first substrate, and in a second area, orthographic projections of at least part of the second sub-support column on the substrate are overlapped with orthographic projections of the touch signal lines on the first substrate, and the supporting component and the touch signal lines are mutually insulated.

The invention discloses an automatic gain control method for audio frequency, which presets the initial gain value. The method comprises: check whether the present frame input signal is mute frame signal; if it is a mute frame signal, directly output the present frame input signal; otherwise, process the present frame input signal in an amplification gain device and then output it. Besides, the invention also discloses an automatic gain control device for audio signal. The method and device disclosed in the invention, which do not amplify the background noise and do not produce signal distortion, are able to realize the automatic control of output power in hearing aid under the condition of ensuring the speech resolution ability.

The invention discloses a low-dose X-ray CT image reconstruction method. The low-dose X-ray CT image reconstruction method sequentially comprises the steps that (1) imaging system parameters of CT equipment and projection data in low-dose CT scanning protocol are obtained, (2) the projection data variance on each data point is calculated according to the projection data obtained in the step (1) and the position of each data point and the number of all the data points are showed, (3) a projection data recovery model is established according to the projection data variances estimated in the step (2), (4) solving is carried out according to the projection data obtained in the step (1) by means of the projection data recovery model established in the step (3) to obtained recovered projection data, and (5) image reconstruction is carried out according to the recovered projection data obtained in the step (4) to obtain a reconstructed CT image. The low-dose X-ray CT image reconstruction method has the advantages that noise and artifact of the image can be greatly reduced, the resolution ratio of the original image can be well kept, and high-quality reconstruction of the CT image is finally achieved.

The invention provides an image processing method, an image processing apparatus, and a terminal. The method comprises the following steps: obtaining image data of N images, wherein N is an integer greater than 1; respectively performing pixel merging processing based on a first scale on the image data of the N images to obtain image data of a first image obtained by merging the N images; and performing upsampling on the image data of the first image obtained by merging the N images based on a super-resolution reconstruction algorithm to obtain image data of a first super-resolution image, wherein the resolution of the first super-resolution image is greater than the resolution of the image obtained by merging the N images. By adoption of the scheme, a better dynamic range and more detailscan be obtained on the premise of ensuring the image resolution.

The invention provides a multispectral collection system, which comprises an imaging objective lens device, a sampling mask device, a prismatic decomposition device, a photographing device and a spectrum collection calibration device, wherein the imaging objective lens device is used for converging and transmitting scene light rays once to generate transmission light rays and comprises a first lens and a second lens; the sampling mask device is positioned at an imaging plane between the first lens and the second lens and is used for performing space sampling on a scene; the prismatic decomposition device is positioned at the rear end of the imaging objective lens device and is used for performing chromatic dispersion on the transmission light rays which pass through the imaging objective lens device to form spectra on a plurality of wavelengths; the photographing device is positioned at the rear end of the prismatic decomposition device and is used for imaging the spectra on a plurality of wavelengths to generate an image to be calibrated; and the spectrum collection calibration device is positioned at the rear end of the photographing device and is used for calibrating and correcting sampling points of the image to be calibrated to acquire spectrum information of the sampling points of the scene. The invention also provides a multispectral collection method. By the multispectral collection system, high-resolution multispectral data acquisition with large range of depth of field and luminous flux can be realized.

An MEMS accelerometer comprises a substrate, an acceleration sensitive chip and an integrated circuit, wherein the acceleration sensitive chip and the integrated circuit are arranged on the substrate, a temperature sensor is further arranged on the substrate and is adjacent to the acceleration sensitive chip, and the acceleration sensitive chip comprises a plurality of measurement bodies. Each measurement body comprises a mass block and a comb tooth structure, wherein the comb tooth structure comprises movable comb teeth extending from the mass block and fixed comb teeth, the movable comb teeth and the fixed comb teeth are arranged at intervals and form a difference detection capacitor, the comb tooth structure is divided into multiple groups, the multiple groups of comb teeth structures are divided into a first part and a second part, the variation direction of movable intervals of the comb tooth structure of the first part and the variation direction of movable intervals of the comb tooth structure of the second part are opposite, and one group of the comb tooth structure and one group of the comb tooth structure of the second part respectively output an electric signal. The MEMS accelerometer resolves the problems that a traditional MEMS accelerometer is large in temperature factor and large in temperature hysteresis.

The invention provides an image registration method and an image registration device based on a cascade structure. The method comprises the steps: carrying out bilateral matching processing on a characteristic point of a reference image and that of an input image to obtain a first coarse matching group; eliminating matched pairs in the first coarse matching group according to a predetermined confidence coefficient to obtain a second coarse matching group, wherein the similarity of the matched pairs eliminated from the first coarse matching group is lower than the predetermined confidence coefficient; eliminating the matched pairs in the second coarse matching group according to a predetermined threshold value to obtain a third coarse matching group, wherein the gray distance or the space distance between any two matched pairs in the matched pairs eliminated from the second coarse matching group is greater than the predetermined threshold value; processing the matched pairs in the third coarse matching group by adopting an RANSAC (RANdom SAmple Consensus) algorithm; and aligning the reference image with the input image spatially according to a transformational matrix so as to obtain an output image.

The embodiment of the invention, which relates to the technical field of camera testing, discloses a camera testing system and method, a terminal and a computer-readable storage medium. The testing system is composed of a control unit, a test light box, a test stand arranged in the test light box, and a test picture card arranged in the test light box. The test stand is used for fixing a to-be-tested terminal and being connected to a camera of the to-be-tested terminal. The control unit in communication connection with the test stand and the camera is used for controlling the test stand to move freely in the test light box, triggering the camera to carry out shooting after the test stand reaches a designated position, and acquiring image information shot by the camera. With the camera testing system and method, the terminal and the computer-readable storage medium, while the need of testing the camera of the terminal is met, the testing requirements of the resolution, ambient light source types, and picture quality and the like are guaranteed and the testing costs are lowered and the economic and practical effects are good.

The invention relates to an electron microscope identification method for residual oil in light reservoir nanopores and a device for executing the method. The method comprises the following steps: obtaining an electron microscope sample; processing the electron microscope sample to obtain a flat section with exposed micro-nanopores; obtaining an back scattering image without an electric charge phenomenon on the flat section with the exposed the micro-nanopores, wherein the back scattering image without the electric charge phenomenon displays the micro-nanopores and distribution of the protogenetic organic matter; on the same imaging position, obtaining an back scattering image of residual oil charges, which displays the micro-nanopores and distribution of the protogenetic organic matter; respectively extracting the pores from the back scattering image without the electric charge phenomenon and the back scattering image of the residual oil charges to obtain distribution of the residual oil in a three-dimensional porous space, so as to realize identification of residual oil in the light reservoir nanopores.

The invention relates to a variable-focus 3-D capsule endoscope system based on a liquid lens. Design is conducted by adding the liquid variable-focus lens to a capsule endoscope optical system, focusing of the system is achieved through zooming of the liquid lens, and therefore clear images of a shot target can be obtained at any position. Sequences of the obtained clear images are processed and calculated, more accurate three-dimensional information of the shot target is obtained, and therefore the 3-D imaging or three-dimensional ranging function is achieved for the shot object, and doctors can better judge diseased parts and lesion degrees. Clearer images of human body alimentary canals can be obtained, the calculation accuracy is improved, and the three-dimensional information of the shot target is more accurately obtained. The capsule endoscope diagnosis accuracy and the work efficiency of the doctors can be effectively improved.