49 results about "Imaging algorithm" patented technology

The application relates to a method for detecting an obstacle in a parking space and a storage medium, the method comprising: collecting a parking space image; splitting the parking space image into a plurality of sub-images; determining texture features and color features of each of the plurality of sub-images; determining texture similarity of adjacent two of the plurality of sub-images with respect to the texture features and color similarity with respect to the color features; forming a difference coefficient of the parking space image according to statistical information of the texture similarity and the color similarity; and judging whether an obstacle exists in the parking space according to the difference coefficient. The method can determine whether an obstacle exists in the parking space according to an image algorithm, so as to guarantee parking safety.

The application discloses a kind of acoustic-electric tomography image imaging algorithm based on two-point gradient iterative method, and the innovation is to improve the calculation efficiency by two-point gradient iterative method, and the accuracy of image reconstruction is improved by the selection of regularization term.The core of the application is to use the characteristics of acoustic-electric hybrid imaging, that is, when the boundary detection current is applied, the power density data is obtained by using ultrasonic disturbance, so as to obtain the internal data of biological tissue.At the same time, according to the different characteristics of different biological tissue conductivity, different regularization terms are introduced, two-point gradient iterative method is established, and high-resolution reconstruction of biological tissue conductivity is realized.Compared with the traditional acoustic-electric imaging method, the application reduces the calculation and storage cost of multiple observation currents, improves the imaging efficiency.And the iteration process is modified by using a special regularization term, so that the boundaries or geometric characteristics of different tissues can be accurately determined, and the imaging accuracy is improved.

This disclosure describes systems, methods, and devices related to video post-processing using a single local directional pattern (LDP) for multiple post-processing steps. A method may include identifying video received by a first device from a second device and decoded by the first device; generating a LDP of a video frame of the decoded video; detecting, based on the LDP and the video frame input to a blurred region detection algorithm, a blurred region and a non-blurred region of the video frame; applying, based on the LDP and the video frame input to a super resolution algorithm, super resolution on the non-blurred region of the video frame without applying the super resolution to the blurred region; and generating, based on the LDP and the video frame input to a blended image algorithm, a blended image of a low-resolution image of the video frame and a high-resolution image of the video frame.

The invention provides a depth domain velocity modeling execution engine design method and device and a medium, and belongs to the field of seismic exploration data processing. The method comprises the following steps: constructing a process of depth domain velocity modeling; carrying out feasibility verification on the constructed process of depth domain velocity modeling; and according to the constructed depth domain speed modeling process, iteratively executing a task for multiple times to complete depth domain speed modeling. According to the method, automatic scheduling execution of the algorithm modules in the process and iterative updating of the whole-process speed model are completed according to the depth domain speed modeling process which is self-defined and constructed by a user, deep integration and fusion of various high-end migration imaging algorithm modules are realized, construction of the depth domain speed model is conveniently and efficiently realized, and the speed modeling efficiency is improved. And revolutionary technical progress is brought to the fields of seismic exploration and the like.

The application discloses an electrical tomography imaging method based on electromagnetic field momentum and a signal detection device thereof, and introduces ET-EMM by taking a current source or a charge source excitation detection mode as an example. The steps of the ET-EMM method are as follows: under the excitation of the current source or the charge source, the boundary value problem of a measurement area is solved to obtain the spatial distribution F of electromagnetic fields of an imaging area; the ET-EMM sensitivity matrix is constructed according to the spatial distribution F of the electromagnetic fields; the electric field intensity or the magnetic induction intensity of a receiving unit of an electromagnetic sensor is obtained through a signal detection part; and finally, image reconstruction is carried out by using the ET-EMM sensitivity matrix. The signal detection device of the ET-EMM mainly comprises a signal generation part, an electromagnetic sensor, a signal detection part, a signal amplification part and a signal acquisition and imaging algorithm part. The ET-EMM method disclosed by the application obtains higher-order tensor information than the measurement parameters of the electrical tomography, and further improves the resolution of the electromagnetic characteristic parameter distribution image.

The application discloses a curved surface array implicit aperture expansion imaging method based on sparse enhanced iterative adaptive processing, belongs to the technical field of array signal processing and beam imaging, and particularly relates to a high-resolution imaging algorithm for a non-planar (curved surface) conformal array. The core realizes accurate beam focusing, sidelobe suppression and robust beam pointing through virtual element expansion and iterative adaptive (IAA) beam shaping technology, and can be widely adapted to various systems such as radars, sonars and radio positioning systems.

Provided in the embodiments of the present application are a formation imaging system, a formation imaging method, a device and a medium. The system performs formation imaging on the basis of a reverse time migration (RTM) formation imaging algorithm. The system comprises: at least one group of nodes, and a CXL 2.0 memory pool in communication with the at least one group of nodes, wherein the at least one group of nodes are used for inputting computation process data into the CXL 2.0 memory pool in a forward propagation simulation stage of the RTM formation imaging algorithm, and reading the computation process data from the CXL 2.0 memory pool in a backward propagation simulation stage of the RTM formation imaging algorithm; and the CXL 2.0 memory pool is used for dynamically allocating memory to the at least one group of nodes. By means of the embodiments of the present application, the dynamic requirements of the RTM formation imaging algorithm in a cluster for memory can be met, and the effective management of a public memory pool in the cluster is realized, thereby reducing the consumption of hardware resources.

This invention proposes an imaging method and system based on the coordinated extraction and correction of dual harmonics with high signal-to-noise ratio using narrowband magnetic particles. The method includes: Step S1: Constructing a field-free point formed by a pair of permanent magnets; Step S2: Injecting a low-frequency, high-amplitude current into a coil to form a driving field; Step S3: Injecting a high-frequency, low-amplitude current into a coil to form an excitation field; Step S4: Constructing a subsequent signal selection and processing module; Step S5: Constructing a precise analog-to-digital conversion module; Step S6: Designing a signal imaging algorithm. This invention also includes: a pair of permanent magnets for establishing a magnetic field; a coil, into which current is injected to form the driving and excitation fields; a subsequent signal selection and processing module, including a transmission chain and a detection signal chain; an analog-to-digital conversion module, including an AD module and an FPGA module; and a signal imaging algorithm module, including imaging using the X-space algorithm.

This invention discloses an intelligent target detection and real-world annotation system based on acoustic-optical imaging, comprising: a multimodal data acquisition module for acquiring acoustic wave data and visible light image data; a data processing circuit module for preprocessing the microphone acoustic wave data and visible light image data; an intelligent target detection and localization module that uses a deep learning algorithm to detect targets of interest in the visible light image, while simultaneously matching the acoustic imaging algorithm module to perform sound source localization and sound field imaging of the targets of interest, and then fusing the acoustic imaging results with the target detection results in the visible light image to achieve target detection and localization; and a real-world annotation module that uses a projector or matrix lighting to generate projected content based on the detection results, enabling interactive annotation of the observation space. This invention can achieve multimodal detection and localization of targets of interest in different scenarios, and is particularly suitable for occluded and hidden targets. Furthermore, by adding an annotation projection module, realistic human-computer interaction can be achieved.

This invention discloses a fast autofocusing method for layered imaging based on vortex edge enhancement and selective linear fitting, comprising: acquiring the initial complex amplitude of an object; propagating the initial complex amplitude to a virtual Fresnel zone plate, modulating it to generate a hologram, extracting edge information during the backpropagation of the hologram, locating the region of interest (ROI), and calculating the sharpness bias within this ROI, storing it along with the corresponding axial distance in a historical dataset; during iterative iteration, substituting the updated distance parameters for each iteration of the layered imaging algorithm, classifying all data points in the historical dataset into linear or saturated intervals, performing least-squares fitting on data points classified as linear intervals, outputting the intercept and slope of the fitting function, and calculating new axial distances based on the current data point classification using different strategies, before proceeding to the next iteration. This invention enables high-precision and high-efficiency automatic calibration in a single preprocessing step.

This invention discloses an AI training image algorithm for CT scans, comprising: inputting an initial image Image, wherein the initial image is a clinical routine CT noise reduction image or a medical model CT image; performing orthographic projection processing on the initial image to generate 180° multi-channel attenuation values; determining the target photon count according to the required noise level, obtaining the maximum projection value among the multi-channel attenuation values, converting the attenuation values ​​of each channel into photon counts, and generating Poisson noise and Gaussian noise; performing modulation processing on the Poisson noise and Gaussian noise corresponding to each channel to generate the comprehensive noise of that channel, adding the comprehensive noise of each channel to the photon count of the corresponding channel to obtain the noisy photon count; converting the noisy photon count of each channel to the projection domain to obtain the noisy attenuation value, and performing back projection processing on the noisy attenuation value to generate a simulated CT noise image for AI training, which can reduce clinical image noise, improve image diagnostic reliability and equipment performance.

This invention discloses a method for locating the void between the surface layer and base layer of an asphalt pavement, belonging to the field of pavement inspection technology. The method includes: scanning the pavement with ground-penetrating radar to obtain radar data; performing horizontal filtering on the radar data; and applying an improved back projection imaging algorithm to the horizontally filtered radar data to obtain the true location of the void. This invention performs horizontal filtering on the acquired radar data to reduce the influence of reflected waves from the surface layer and base layer interface on void identification, minimizing the influence of direct waves and interface reflections. This invention only superimposes radar signals adjacent to grid points, greatly reducing the computational load, and can accurately locate the void position.

This invention provides a broadband atomic terahertz detection method and apparatus, relating to the field of terahertz detection and imaging technology. The method includes: generating a time-coded sequence using an arbitrary waveform generator; controlling a microwave switch to perform orthogonal coding modulation on a terahertz field to generate a terahertz differential signal carrying coded information; applying the terahertz differential signal to Rydberg atoms in an atomic gas cell to convert terahertz information into changes in the transmitted light intensity of the probe light, and converting the optical signal into an electrical signal; synchronously demodulating and acquiring the electrical signal to generate a measurement vector; and reconstructing the original terahertz time-domain waveform from the measurement vector using a compressed sensing reconstruction algorithm. This invention, by combining superheterodyne detection and time-domain computational imaging algorithms, overcomes the physical limitations of the bandwidth of atomic detectors, achieving rapid and high-resolution detection of broadband terahertz signals while maintaining their high sensitivity advantages.

This disclosure provides a method and apparatus for detecting the oil level in a compressor. The method includes: acquiring real-time operating parameters of the compressor and an image of the oil level to be measured; matching the timestamps corresponding to the real-time operating parameters and the image of the oil level to be measured; inputting the real-time operating parameters into an oil level height regression prediction model to output a confidence interval for the oil level height of the compressor; determining all candidate oil level heights of the compressor based on the image of the oil level to be measured using an image algorithm; and determining the current oil level height detection result of the compressor based on the confidence interval and the candidate oil level heights. This disclosure utilizes a joint decision-making mechanism based on multimodal data to achieve linked analysis of the oil level state based on time-series data of real-time operating parameters and oil level image data, thereby fully utilizing multimodal data during compressor operation to infer the oil level state and improving the prediction accuracy of compressor oil level detection.

This invention belongs to the field of external insulation equipment inspection technology, and discloses a back projection imaging method, system, equipment, and medium for defect detection of external insulation equipment in power systems. It addresses the problem of low imaging efficiency caused by the large amount of repetitive calculation of electromagnetic wave propagation paths in existing back projection imaging algorithms. The method utilizes the axisymmetric structural characteristics of the external insulation equipment. First, a reference detection orientation is selected to calculate its electromagnetic wave propagation path distribution. Then, this reference path distribution is rotated to other detection orientations to obtain the path distribution to be interpolated. Next, an inverse distance weighted interpolation method is used to accurately calculate the target pixels. Finally, the above steps are repeated to obtain path data for all detection orientations for imaging. This invention significantly reduces the repetitive calculation of electromagnetic wave propagation paths through parameter reuse, greatly improving the efficiency of back projection imaging while ensuring imaging accuracy, and meeting the needs of rapid on-site inspection of external insulation equipment in power systems.

This invention proposes a single-station localization method for non-line-of-sight targets based on angle-constrained back projection, aiming to solve the problems of low localization accuracy and low noise tolerance for non-line-of-sight targets in complex environments. Applied to the field of non-line-of-sight target detection technology, this method addresses the limitations of traditional back projection imaging algorithms. These algorithms are constrained by underlying time delay compensation and coherent accumulation mechanisms, requiring large-span physical spatial view diversity to overcome the azimuth resolution bottleneck. Furthermore, in complex environments, they are highly susceptible to background clutter and multipath sidelobe interference, and face the challenge of blind correlation due to the unknown mapping relationship between multipath signals and physical reflective surfaces. This method first introduces an angle-of-arrival constraint to limit the spatial imaging range of the back projection algorithm, obtaining a grid energy spectrum for multipath energy focusing. Subsequently, by executing a local extremum optimization strategy in the image domain, clutter and multipath sidelobe interference in the energy spectrum are accurately removed, and spatial discrete imaging feature points with strong multipath interference are extracted with high confidence. Finally, the system introduces a building layout map as a spatial topological constraint prior. Based on the aforementioned feature points, it constructs and traverses the association assumptions between multipaths and physical walls to perform a mirror-inverse mapping of the propagation link, and optimizes global coherence energy within the solution space, thereby accurately resolving the final physical location coordinates of non-line-of-sight targets. This effectively overcomes the challenges of blind association and limited physical aperture.

The application discloses a rapeseed suitable fertilization amount test screening method, and comprises the following steps: collecting chlorophyll fluorescence parameters of rapeseed canopy in each test area, screening test areas with abnormal chlorophyll fluorescence parameters according to the chlorophyll fluorescence parameters, and generating a rapeseed canopy physiological state early warning list; positioning acoustic sensor nodes based on the rapeseed canopy physiological state early warning list, generating an acoustic sensor node group, and scanning soil in a rapeseed root area according to the acoustic sensor node group to obtain a soil medium acoustic characteristic data set; converting the soil medium acoustic characteristic data set into an acoustic wave velocity distribution diagram and a signal absorption coefficient diagram by means of an acoustic imaging algorithm, and performing spatial registration and data fusion to generate a rapeseed root area soil acoustic characteristic profile diagram. The application realizes nondestructive visual characterization of the root area soil structure and root system morphology, thereby providing accurate root system spatial distribution and physical characteristic insight.

The application particularly relates to a kind of near distance multi-target fast deformation monitoring methods based on small aperture MIMO radar, high azimuth resolution two-dimensional imaging of multiple azimuth adjacent targets is realized by discrete fourier transform combined with MUSIC algorithm, and accurate positioning of multi-target is completed;Secondly, the mutual interference between target side lobe is suppressed using APES filter, and based on the target distance and azimuth information obtained in the previous step, the accurate recovery of target phase is realized;Finally, the time difference interference method is used to accurately estimate the small deformation of target visual distance.The space smoothing (FBSS) idea is introduced into the two-dimensional imaging algorithm, which effectively suppresses the coherent signal interference and realizes high azimuth resolution two-dimensional imaging in complex deformation monitoring scene.

The application is based on the ultrasonic phased array wind power blade interlayer bonding interface fitting method and system, and belongs to the nondestructive testing technical field, comprising: configuring a detection system and initializing parameters; performing ultrasonic phased array full matrix data acquisition and preprocessing on the blade bonding part; calculating the ultrasonic wave propagation time delay based on the multilayer medium sound velocity distribution model and the ray tracing method, and generating a high-resolution glue layer-blade layer bonding interface image by using a full focusing imaging algorithm; adopting a multi-scale edge detection algorithm combining a Canny operator and a Zernike matrix to accurately extract a discrete feature point set of the bonding interface from the image; based on the feature point set, a continuous contour curve of the bonding interface is obtained by using a cubic polynomial curve fitting; and finally, the fitting curve is superimposed on the imaging result for visualization. The application effectively solves the imaging distortion problem caused by the sound velocity difference in the heterogeneous multilayer structure, and significantly improves the detection reliability of the wind power blade interlayer bonding quality.

This invention discloses a mid-orbit curve SAR frequency domain imaging method based on spatially varied azimuth preprocessing. This invention relates to an imaging algorithm that, under long synthetic aperture time and slant-look observation geometry conditions, suppresses the two-dimensional spatial variation of range cell migration (RCM) through improved azimuth resampling technology, thereby achieving high-precision focusing. This invention is achieved through the following technical solution: Step 1, constructing a two-dimensional spatially varied model of the higher-order slant range history and its parameters for MEO SAR. Step 2, establishing an improved azimuth resampling model based on range gate dependence. Step 3, solving for the optimal azimuth resampling coefficients. Step 4, performing improved azimuth resampling and imaging focusing processing.

The application relates to a method for intercepting an effective imaging scene of large-oblique SAR imaging by radar technical indexes, and belongs to the technical field of large-oblique SAR imaging. According to technical index requirements of a radar signal processor, such as a signal bandwidth, a sampling rate, a pulse accumulation time and the like, which are set by a project, the following specific requirements are calculated: an imaging scene center distance, an imaging width and a resolution in a distance direction and an azimuth direction of an imaging scene in a ground plane, and the like; according to a conversion relationship of oblique plane ground plane coordinate axes, the following are calculated: a distance dimension point number and an azimuth direction pulse pressure accumulation point number of a two-dimensional imaging scene in an oblique plane, so that effective data in data recording is retained, redundant data is removed, time consumption of real-time processing of an imaging algorithm is shortened, and design requirements are met.

This utility model relates to the field of mechanism gear technology and discloses an automated inspection device for mechanism gears, including a housing. The device uses a motor fixed to the inner wall of the housing. When the motor runs, it drives the gear to rotate via a connecting rod. The gear meshes with a gear ring, which in turn drives a transparent rotating disk to rotate. The transparent rotating disk slides along the outer wall of the inner groove via a rotating block, allowing the product to move on the disk. A support column supports a visual inspection camera array, which operates with a layout of two cameras (top and bottom) and three cameras (horizontal). The top camera vertically captures the top surface of the gear, the bottom camera captures the bottom surface of the gear through the transparent rotating disk, and the three horizontal cameras focus on the surface finish and inter-tooth gaps, respectively. Industrial image algorithms are used to extract edges and compare features from images at various angles for accurate identification, thereby accurately inspecting the product, improving inspection efficiency, and ensuring the stability of inspection quality.

The application relates to the technical field of image processing, and discloses a rain day image processing system based on an intelligent image algorithm, which comprises a vanguard data acquisition module, a vanguard data processing module, a result storage module, a rain day image acquisition module, a humidity information acquisition module, a core processing module and a real-time feedback module; the vanguard data acquisition module is used for acquiring normal image data in a recent period of time, selecting image data meeting a condition from the normal image data, and acquiring humidity data corresponding to the image data meeting the condition; the vanguard data is analyzed and processed to obtain a brightness threshold value and a humidity threshold value; when rain day image processing is carried out, raindrops in the image can be accurately recognized by using a three-dimensional coordinate system and an RGB value; the raindrops in the image are removed by using a pixel point filling method; and water vapor in the image is removed by using the humidity threshold value, so that the whole image becomes clearer.

The application discloses a kind of based on electromagnetic field momentum's electrical capacitance tomography method and gradient signal detection device, the method is based on electromagnetic field momentum reciprocity theorem, reconstructs dielectric constant by the gradient signal of capacitance.It steps are, first, solve boundary value problem under voltage source excitation, construct sensitivity matrix or the gradient of sensitivity matrix;Second, construct ECT-EMM linear equation under suitable coordinate system;Then, obtain the gradient signal of sensor receiving electrode's capacitance by signal detection part;Again, reconstruct dielectric constant distribution using sensitivity matrix or the gradient of sensitivity matrix, coordinate system related coefficient vector and the gradient signal of capacitance.It gradient signal detection device includes: signal generating part, sensor, signal detection part, signal amplification part and signal acquisition and imaging algorithm part.The imaging method proposed in the application has the advantages of high imaging resolution and can meet the needs of different industrial environments.

The application discloses an ultrasonic imaging method, device, equipment and storage medium. The method comprises the following steps: scanning a target object based on a focused wave signal or a plane wave signal to obtain echo data; performing data transformation on the echo data based on a preset transformation algorithm to obtain first image data, so that sparse sampling data is restored to image data consistent with a probe shape; performing signal level processing and image level processing on the first image data to obtain second image data, so that the focused wave and the plane wave can be simultaneously compatible with the same ultrasonic system, thereby enabling the ultrasonic system to quickly switch the focused wave and the plane wave without two sets of imaging algorithms; and finally, performing mask covering on the second image data based on preset mask data to obtain a target ultrasonic image, so as to eliminate invalid information of the second image data, and enable the focused wave and the plane wave to be simultaneously well compatible with the same ultrasonic system.

This invention relates to the field of image design technology and discloses an image style design method and system based on artificial intelligence. The method includes: extracting initial fields of prompt words from input text; generating high-dimensional vectors based on the initial fields; sampling in a probability distribution space based on the high-dimensional vectors; determining supplementary fields of prompt words based on predicted word occurrence probabilities; combining the initial and supplementary fields to determine preliminary text; performing repetition analysis on the target text to determine the order of each keyword; training a multimodal large language model using a diffusion model and a masking algorithm to determine the target multimodal large language model; determining the visual image of the target text based on the order and the target multimodal large language model; and processing the visual image using a preset image algorithm to determine the target visual image. This invention improves the reliability of cover design by generating image style images through artificial intelligence.

The application relates to a water antenna detection method for an unmanned ship sensing system, and the features include the following steps: S1) establishing a water antenna detection convolutional neural network structure, wherein the water antenna detection convolutional neural network structure comprises a backbone network for feature extraction and a branch prediction output head network; S2) training the water antenna detection convolutional neural network structure established in the step S1 by using a training data set; S3) when the training reaches an iteration number Q, fixing and saving all weight values w and bias values b in the network to obtain a trained convolutional neural network model W; S4) storing all (x, y, theta) in a container allpoints; and S5) obtaining a water antenna image coordinate straight line equation y=k*x+b. The water antenna detection method can avoid the low image segmentation precision of a traditional image algorithm, cannot obtain accurate water antenna detection results, has high running speed and high reliability.

This invention relates to the field of nondestructive testing (NDT) technology, specifically disclosing an ultrasonic NDT method based on full-matrix data transfer and phase coherence, used to detect internal damage defects in planar multilayer structures in equipment such as pressure vessels. This method utilizes the time-domain echo signal received by an ultrasonic phased array body wave probe, combined with frequency-domain phase shift transfer of the full-matrix dataset to achieve rapid extrapolation of the wave field, improving the algorithm's time efficiency by avoiding interface refraction operations of the time-domain signal. Based on the explosion reflection model, the time-domain dataset of the target region is extracted, and combined with a phase-coherence weighted full-focusing imaging algorithm, accurate localization of internal damage within the target layer of the planar multilayer structure is achieved. Experimental results show that this invention has high accuracy and resolution in defect localization imaging of planar multilayer structures, and is suitable for the inspection and maintenance of planar multilayer walls in equipment such as pressure vessels.