48 results about "Noise map" patented technology

The invention discloses a remote sensing image super-resolution reconstruction method and device, a storage medium and equipment, relates to the technical field of computer vision, and can solve the technical problem that the remote sensing image super-resolution reconstruction effect is poor. The method comprises the following steps: acquiring a sample remote sensing image and multi-modal auxiliary data, extracting spectral features, textural features and geometric features of the pre-processed multi-modal auxiliary data, and fusing to generate a condition feature tensor; performing forward diffusion on the sample remote sensing image to generate a noise-containing image sequence; in the process of training the UNet diffusion network by using the noisy image sequence, injecting the conditional feature tensor into preset levels of an encoder and a decoder of the UNet diffusion network, and optimizing network parameters through multi-objective loss to obtain a trained UNet diffusion network; and obtaining a to-be-reconstructed noise map, performing noise prediction on the noise map based on the trained UNet diffusion network to obtain target noise, performing reverse sampling on the noise map, and generating a target reconstructed image through iterative denoising.

The invention provides a marine acoustic environment monitoring system based on a noise map. The marine acoustic environment monitoring system comprises a buoy network module, a data fusion processing module, a noise map generation module, a noise source identification module and an edge response module, the buoy network module serves as a sensing tail end, multi-source information is gathered to the data fusion processing module through the first data bus for standardization processing, and then the noise map generation module is driven through the second data bus to construct a three-dimensional sound field model. The output of the model is transmitted to a noise source identification module through a third data bus for intelligent analysis and traceability; the alarm decision of the noise source identification module is output to the edge response module, and the edge response module dynamically regulates and controls the communication and acquisition strategy of each buoy node through a distributed control circuit to form a complete closed loop. According to the invention, high-precision and automatic monitoring of the marine acoustic environment is realized, the problems of limited monitoring range and single data dimension of the traditional means are solved, and the reliability and emergency capability of the system are improved.

The invention discloses a colored shaded relief map generation method and system fusing land coverage information. The method comprises the steps of obtaining DEM data and land coverage data of a target area, taking the preprocessed DEM data as DEM condition information, and taking the preprocessed land coverage data as land coverage condition information; the method comprises the following steps: inputting a randomly generated noise image and DEM condition information into a backbone network, inputting the noise image, the DEM condition information and land coverage condition information into a ControlNet fine tuning module to obtain a correction signal, inputting the correction signal into the backbone network to obtain predicted noise, and performing reverse denoising based on the predicted noise to obtain a colored shaded relief map. According to the method, the semantic correctness of the generated color can be ensured, and the generated colored shaded relief image better conforms to the visual effect of a real natural landscape.

The embodiment of the invention relates to an image processing method and device, equipment and a medium. The method comprises the steps that a first image and a first noise map are acquired; wherein the first image comprises a target object, and the transparency of the target object in the first image is different from the transparency of an area outside the edge of the target object; performing fuzzy processing on the first image to obtain a fuzzy image corresponding to the first image; generating a second image corresponding to the first image based on the transparency information of the blurred image and the first noise map; wherein the edge of the target object in the second image is obtained by acting the first noise map on the edge of the target object in the first image based on the transparency information of the blurred image. According to the invention, special effect processing on the object edge in the image can be efficiently and conveniently realized, and the realization efficiency of the burr special effect is effectively improved.

The control device (20) of the present application includes: an image acquisition unit (203) that acquires a radiograph of a subject (F) irradiated with radiation and photographed through the subject (F); a noise map generation unit (204) that derives an evaluation value based on relationship data representing a relationship between a pixel value and an evaluation noise value, from a pixel value of each pixel of the radiograph, and generates a noise map that corresponds to the derived evaluation value and each pixel of the radiograph; and a processing unit (205) that inputs the radiograph and the noise map into a learning completed model (207) constructed by machine learning in advance, and performs image processing for removing noise from the radiograph.

A method and apparatus with noise consideration are provided. The method includes generating, using a noise model, a non-normal noise map corresponding to a noise of an input image, and generating an enhanced image of the input image by implementing an image enhancement model based on the input image and the non-normal noise map, where the noise of the input image follows a non-normal distribution.

The present application relates to a kind of computing method for realizing wide range dynamic noise map system fast update, comprising the following steps: using the data information of system database, based on existing day and night period and city road grade, simplify the structure of traditional road sound level mechanism model;Using big data technology, using hybrid modeling method, regression each grade road sound source (day and night) sound level contribution amount prediction model;Road noise fast calculation method based on sound level contribution amount prediction model is constructed, the fast update of dynamic noise map is realized.Compared with prior art, while ensuring the calculation precision, the operation and update efficiency of noise map system is greatly improved, and the technical constraints of foreign core computing software are broken, the localization of noise map core computing technology is realized, which provides technical support for wide range deep development of noise map.

To provide a data processing method and a data processing system capable of improving accuracy of living body detection in a vehicle.SOLUTION: A data processing method for processing data collected by a millimeter-wave radar, comprising: collecting in-vehicle actual measurement data; and removing noise in the in-vehicle actual measurement data based on a noise map, wherein the noise map is generated based on in-vehicle environment data collected when it is determined that no living body exists in a vehicle.SELECTED DRAWING: Figure 3

The invention provides an image processing method and device and electronic equipment, and the method comprises the steps: obtaining an initial image, inputting the initial image into a pre-trained stable diffusion model, and generating a noise graph through the stable diffusion model; based on a self-attention mechanism, generating disturbance data corresponding to the noise map, and determining target prediction noise through a stable diffusion model and the disturbance data; and performing denoising processing on the noise map based on the target prediction noise to obtain a target image. According to the mode, the structural features of the original image are stored in the image generated by the stable diffusion model, and the image quality of the image generated by the stable diffusion model is improved.

Embodiments of the present application provide a scene special effect processing method and device, electronic equipment and storage medium, and relate to the technical field of image processing, the method comprises: acquiring scene data corresponding to a game scene, the scene data comprising a scene model and map data, the map data comprising a special effect mask map and a noise map; acquiring color control information for adjusting a gradient color, performing gradient processing on the color control information and the texture coordinates of the scene model to obtain a gradient color map corresponding to the scene model; acquiring noise control parameters for the noise map, performing flow control on the noise map according to the noise control parameters to obtain a flow noise map corresponding to the noise map; performing vertex displacement according to the noise map to obtain offset information corresponding to the scene model; and rendering and outputting the scene model according to the gradient color map, the flow noise map and the offset information, so that the game scene is more realistic and the visual effect is more rich.

This invention discloses a training method, system, and device for an underwater image enhancement model, relating to the field of image enhancement technology. The method includes: extracting the natural light field map and detail map of an underwater image; merging the detail map with the natural light field map after Gaussian filtering to obtain a detail degradation map; determining non-offset color channels in the green and blue channels of the natural light field map, performing color compensation on the red and non-offset color channels, and merging the color-compensated natural light field map with the detail map to obtain a color compensation map; reducing the brightness of the underwater image to obtain a low-brightness map; adding Gaussian noise to the underwater image to obtain a noise map; and using the detail degradation map, low-brightness map, and noise map as negative samples, and the reference clear image and color compensation map as positive samples, thereby completing the training of the underwater image enhancement model. This significantly improves the color calibration capability and contrast enhancement performance of the underwater image enhancement model for underwater images.

A first aspect of the present disclosure relates to a device for unmixing an image of a sample having fluorescent dyes.SOLUTION: The device is configured to obtain a mixed image of the sample, determine an unmixed image based on the mixed image, determine a noise map image based on the mixed image, determine a signal-to-noise ratio image based on the unmixed image and the noise map image, determine a denoised signal-to-noise ratio image based on the signal-to-noise ratio image, and determine a noise-reduced unmixed image of the sample based on the denoised signal-to-noise ratio image and the noise map image.SELECTED DRAWING: Figure 1

The invention discloses a noise map generation method and device, a storage medium and a program product. Relates to the field of three-dimensional visualization, and the method comprises the steps: obtaining horizontal noise data which comprises the noise data of a plurality of monitoring points in the horizontal direction; according to the method, multiple target acquisition devices are used for noise data acquisition to obtain vertical noise data, and the multiple target acquisition devices comprise noise data acquisition devices deployed at different heights of a target facility; based on the horizontal noise data and the vertical noise data, a three-dimensional noise map is determined through a target network model, and the target network model comprises a neural network model adopting a sound source inversion strategy. Through the method and the device, the problem that the noise map is inaccurate due to the fact that the noise map ignores propagation of noise in the vertical direction in the related technology is solved.

The application discloses a kind of methods for drawing noise map, device and computer readable medium, the method includes the following steps: obtaining detection area, and the detection area is divided into several detection area units;Obtain noise sample, and according to the position of the acquisition equipment of the noise sample, the noise sample is classified into corresponding detection area unit;Based on the noise sample and the position of the detection area unit, obtain the effective noise value of each detection area unit;Based on the effective noise value, render corresponding detection area unit according to predetermined rules, to form noise map.The application effectively solves the technical problems that noise map in the prior art cannot well describe the noise distribution of the whole area, can only show the noise size of detection point, and the visibility is poor, and achieves the beneficial effects of good display effect and good visibility.

The invention discloses an image generation method and device and electronic equipment, and relates to the technical field of image processing, and the method comprises the steps: obtaining a reference image and a detailed text description corresponding to the reference image, and the detailed text description is used for describing the focus of the reference image; performing inversion on the reference image to obtain a noise map containing feature information of the reference image; and matching with the detailed text description, denoising the noise map, and generating a similar image similar to the reference image. The image data of a single scene can be enriched.

The application provides a denoising method and device of an image and electronic equipment. The method comprises the following steps: obtaining image data to be denoised; performing noise component evaluation on the image data to obtain a noise component map corresponding to the image data; performing multi-frequency band decomposition on the noise component map to obtain noise maps of a plurality of frequency bands corresponding to the noise component map; performing different denoising intensity transformations on the noise maps of the plurality of frequency bands based on a plurality of different preset denoising intensities to obtain candidate noise maps of the plurality of different denoising intensities; reconstructing a spatial domain noise map corresponding to the image data based on the candidate noise maps of the plurality of different denoising intensities; and processing the image data according to the spatial domain noise map to obtain a target denoising map corresponding to the image data. Thus, the image data is adjusted in different frequency bands, and the image data is processed in combination with the adjusted spatial domain noise map, so that the target denoising map corresponding to the image data is accurately determined, and the quality of the image is improved.

The invention discloses a wafer defect detection method fusing feature-level and texture-level abnormal synthesis. The method comprises the following steps: obtaining self-adaptive normal features of a normal wafer sample at a normal branch through a frozen feature extractor and a trainable feature adapter; synthesizing global abnormal features by using Gaussian noise in the feature-level abnormal synthesis branch; synthesizing noise maps with different intensities through texture perception defect synthesis in the texture-level anomaly synthesis branch, synthesizing a local anomaly image through texture superposition, and obtaining local anomaly features through a feature extractor and a feature adapter; the three features of the three branches are jointly input into a neural network feature discriminator for end-to-end training; in the reasoning stage, normal branches are used for processing a test image to obtain adaptive features, the adaptive features are input into a neural network feature discriminator to obtain an abnormal score, whether the current test image has defects or not is judged, and abnormal positioning is conducted on the defects. The method provided by the invention can effectively improve the detection capability of weak defects and meet the real-time requirement of industrial production.

The invention discloses a photovoltaic sound barrier layout and power generation potential evaluation method for an expressway, and relates to the technical field of traffic environmental protection, and the method comprises the steps: obtaining spatial geographic data, constructing a model, generating a road noise map, determining a sound barrier layout road section, and calculating the power generation potential of a photovoltaic sound barrier. According to the method, the high-noise sensitive area is accurately identified by combining GIS spatial data analysis and sound environment quality standards through the layout logic driven by the noise map, so that the layout of the sound barrier is changed from experience decision to data drive, the problem that the layout position is not matched with the actual noise influence in the prior art is solved, and meanwhile, the layout accuracy is improved. According to the method, photovoltaic module design and sound barrier structure parameters are integrally considered, available space is quantified through an effective area calculation formula, energy waste caused by disconnection of photovoltaic and sound barrier design is avoided, and a quantifiable technical path is provided for fusion of traffic and new energy.

The invention discloses a speckle elimination method and a laser display system, and belongs to the field of laser display. In order to reduce the influence of speckles, a feature map and a white map in a projection area are shot through a camera, the feature map is used for establishing projection area extraction information to facilitate subsequent extraction of the projection area in an image shot by the camera, and the white map is used for extracting a noise map according to distribution characteristics of pixel values of a local area; and then, according to the noise map and the connected region, judging whether to perform noise adding operation on an output image of the laser display system so as to achieve the purpose of suppressing speckles. According to the invention, the influence of laser speckles on the visual imaging effect can be improved without changing a curtain or an optical structure.

The present disclosure provides a depth map optimization method and device, electronic equipment and storage medium. The method comprises: obtaining a target image and a to-be-optimized depth map corresponding to the target image; using a trained target neural network for depth map optimization to perform feature extraction on the to-be-optimized depth map and the target image to generate a noise map; wherein the noise map is used to represent noise information and / or depth missing information corresponding to the to-be-optimized depth map; using the target neural network and the noise map to perform optimization processing on the to-be-optimized depth map to obtain a target depth map.

The application provides a text-guided image editing method and device, equipment and medium, inputting an input picture, a source prompt word and a target prompt word into a model; the input picture is first added with noise to become a noise picture through an inversion process, and then respectively enters a reconstruction branch and an editing branch; a cross attention map of a reconstruction process of the reconstruction branch is injected into the editing branch to maintain the structure and content of the picture; a contrast attention focusing strategy is adopted in the inversion process; for different degrees of attention leakage, attention is accurately guided to the editing area adaptively; subsequently, a mask-guided mixed sampling technology is used in the editing branch to provide different conditional guidance for the editing area and the non-editing area, so as to ensure that the balance of changes and preservation in the image is integrated in the editing process; only the area needing editing is modified, unnecessary editing of the non-editing area is prevented, and the phenomenon of over-editing is prevented.

A sandstorm prediction method based on multi-source data fusion and conditional diffusion generation defines the sandstorm prediction task as a kind of spatio-temporal sequence prediction problem, the input is the meteorological data at multiple times, and the output is the sandstorm change information at multiple times; the sandstorm data is added with noise according to a noise addition formula to obtain sandstorm data with noise; a multi-source data noise prediction network is trained jointly through satellite cloud images, meteorological reanalysis data, sandstorm data at historical times and a time step t; a full Gaussian noise map, the current time step t x , and satellite cloud images, meteorological reanalysis data and sandstorm data at historical times are input into the multi-source data noise prediction network to predict the noise of the current noise map, the noise at the current time is removed through a denoising formula to obtain sandstorm data with noise at t x‑1 ; this operation is repeated until the t0 time to obtain a prediction result. The present application can accurately predict sandstorms.

A method, apparatus, non-transitory computer readable medium, and system for performing an accelerated denoising process includes obtaining a noise map from a noise distribution. Embodiments then compute, using an image generation model, a denoising vector based on an accelerated denoising trajectory, where the accelerated denoising trajectory accelerates a denoising rate based on a diffusion timestep. Then, embodiments generate, using the image generation model, a synthetic image by denoising the noise map based on the denoising vector.

Method (100) for generating an environmental noise map using a central server (51) and at least one vehicle (52), the method (100) comprising: - Transmitting (110) an initial model to at least one vehicle (52) by the central server (51), wherein the initial model is characteristic of a functional relationship between an input data set and an output data set, wherein the input data set includes a geographical position and wherein the output data set includes an ambient noise level, - Recording (120) at least one noise data set by at least one vehicle (52), the noise data set comprising a geographical position of the vehicle (52) and an ambient noise level at the geographical position, - Updating (130) the initial model for at least one vehicle (52) using at least one noise data set acquired by the vehicle (52), - Transmitting (140) an updated initial model to the central server (51) by at least one vehicle (52), - Aggregating (150) at least one updated initial model into an overall model by the central server (51) and - Generating (160) the environmental noise map using the overall model by the central server (51).

The application discloses a construction area noise map drawing method and system based on crowd-sensing and spatial interpolation, relates to the technical field of construction site noise monitoring, and comprises the following steps: acquiring a measurement distortion threshold when an intelligent terminal device performs noise monitoring; drawing a grid for a worker gathering area to divide the area grid points, collecting noise data of the worker gathering area in a preset target time length based on a crowd-sensing technology, and obtaining an initial noise data matrix; obtaining a noise filling matrix, repairing first invalid point data and second invalid data in the noise filling matrix by using a repair method of a Pearson space-time coefficient-space weight fusion to obtain a repaired noise data matrix; and inputting the plurality of repaired noise data matrices into a noise estimation model to generate a construction area global noise map. The method fully considers the space-time distribution characteristics of noise data in a construction scene, and draws a construction area global noise map.

The application provides a cloud and mist generation method and device and electronic equipment, obtains a 2D noise map and a texture coordinate of each vertex of a plane model; samples the 2D noise map according to the texture coordinate of each vertex to determine an initial vertex offset corresponding to each vertex; obtains an offset factor changing with time for each vertex; further determines a first vertex offset changing with time for each vertex; and renders and generates a target cloud and mist corresponding to the plane model based on the first vertex offset. In this way, the initial vertex offset corresponding to each vertex in the plane model can be determined by sampling the 2D noise map, and the first vertex offset is determined in combination with the offset factor changing with time. Since the first vertex offset of each vertex changes with time, a more dynamic and flexible target cloud and mist can be obtained based on the first vertex offset to match flexible cloud and mist generation requirements.

The invention relates to a method for updating and / or validating a noise map, comprising the steps of: providing (S1) an existing noise map; performing (S2) a sound measurement using a plurality of external microphones (3) of a vehicle (1); determining (S3) a relative position of a sound source (2) of the sound detected during the sound measurement relative to the vehicle (1); determining (S4) an absolute position of the vehicle (1) during the sound measurement and generating a measurement data set by assigning the sound measurement to a determined absolute position of the sound source (2) and / or to the determined absolute position of the vehicle (1); updating (S5) and / or validating the noise map using the measurement data set.

The invention discloses an image super-resolution reconstruction method based on deep reinforcement learning adaptive noise inversion, and the method comprises the steps: firstly constructing a training data set of a low-resolution-high-resolution image pair, constructing an adaptive noise prediction network based on a deep neural network, enabling the network to take a low-resolution image, a time step and a gradient map as joint input, and carrying out the deep reinforcement learning adaptive noise inversion. Outputting a noise map required by inversion of the diffusion model; designing a loss function based on deep reinforcement learning, wherein the loss function is composed of pixel-level loss, perception loss and wavelet loss; carrying out joint training on the noise prediction network and the diffusion model by adopting a training data set and a loss function, and optimizing an adaptive noise inversion strategy to obtain a trained noise prediction network and a trained diffusion model; and performing super-resolution reconstruction on the low-resolution image to be processed by adopting the trained noise prediction network and the diffusion model to obtain a reconstructed high-resolution image.

An image processing method and apparatus, a device and a medium are provided. The method includes: acquiring a first image and a first noise map, where the first image contains a target object, and a transparency of the target object in the first image is different from that of an area outside an edge of the target object; performing blurring processing on the first image, to obtain a blurred image corresponding to the first image; and generating a second image corresponding to the first image based on transparency information of the blurred image and the first noise map, where an edge of a target object in the second image is obtained by applying the first noise map to the edge of the target object in the first image based on the transparency information of the blurred image.

The invention discloses a city noise map generation method based on machine learning, and belongs to the technical field of city sound environment monitoring and noise treatment. According to the method, multi-source open-source data such as remote sensing images, spatial syntax and planning elements are integrated through 75m grids, an urban area is divided into five types such as a hub area and a green area through Gaussian mixture model (GMM) clustering, a noise value is predicted in combination with a random forest (RF) regression model, a feature contribution mechanism is analyzed through SHAP, and finally a noise map is generated. The actual effect of the method on noise prediction is reflected by verifying the urban area trained by the model and selecting a new urban area for testing, and the use requirements are met. According to the method, the existing city element features in the built-up area grid units are recognized, a noise prediction system suitable for a city updating scene is constructed, and accurate improvement and efficient treatment of sound environment quality in the city updating process are achieved.