435 results about "Hue" patented technology

The invention discloses a high-precision image processing method and system based on illumination adaptive compensation, and relates to the technical field of computer vision and image processing, and the method comprises the steps: inputting an original image, and dividing the image into a high-frequency edge layer, an intermediate-frequency texture layer and a low-frequency illumination layer through a multi-scale residual network; acquiring illumination intensity, color temperature and scene categories in real time by using an ambient light sensor and a scene semantic segmentation model, and generating dynamic compensation parameters; carrying out dynamic range expansion on a low-frequency illumination layer based on a physical illumination model, and adjusting the weight of highlight suppression and dark area enhancement through a self-adaptive S-shaped exposure curve; a double-branch generative adversarial network is adopted, noise suppression and super-resolution reconstruction are carried out on the high-frequency layer, and texture detail enhancement is carried out on the intermediate-frequency layer; aligning the data of the depth camera and the infrared sensor with the visible light image through a cross-modal fusion module; and performing tone mapping on the fused image based on human visual characteristics, and outputting an enhanced image with a high dynamic range and reserved details.

The invention relates to the technical field of image recognition, and discloses a food freshness detection method and system based on image recognition, and the method comprises the steps: collecting an initial image, carrying out the white balance correction through a white reference card, and obtaining a corrected image with an accurate color; performing target area segmentation on the corrected image, extracting a food main area and eliminating background interference; pixel sampling is carried out on the main area image, the main area image is converted to Lab, HSV and RGB color spaces, and redness, yellowness, brightness, hue angle and red-green ratio are extracted to form a multi-dimensional color feature set; establishing a difference feature vector based on the feature set and a preset standard sample, and comparing the deviation degrees of the target food and the standard sample on the multi-dimensional color features; and finally, judging the freshness grade of the food according to the deviation degree. According to the invention, lossless and rapid food freshness grading detection is realized.

The invention discloses a transformer substation non-uniform exposure image enhancement method and system based on adaptive fusion, and the method comprises the following steps: obtaining non-uniform exposure video frame images in a transformer substation monitoring video, and carrying out the successive preprocessing of the video frame images; processing the brightness channel graph V by adopting a two-branch parallel step; for the H and S parts of the color channel graph, the saturation of the pre-channel graph is enhanced on the basis of the hue difference; combining the enhanced images of the brightness channel and the color channel with an original channel image, respectively designing a self-adaptive channel weight method for fusion, and converting H, S and V channels after fusion into RGB enhanced images for result output; and establishing an evaluation mechanism to evaluate the quality of the enhanced image, and adjusting parameters of an enhancement algorithm according to an evaluation result. According to the invention, the problem of non-uniform image exposure under the complex illumination condition of the substation is effectively solved, and powerful support is provided for accurately monitoring the state of substation equipment.

The invention relates to the technical field of light control, and discloses a creative and creative light color gradient control method and system, and the method comprises the steps: carrying out the preprocessing of the spatial position information of audiences, generating an audience position sequence, calculating the moving speed, the steering angle and the region density in real time, and generating a light adjustment signal in combination with a preset threshold value; mapping the RGB values to an HSL space to generate an initial color parameter set; adjusting a brightness weight, a hue weight and a saturation weight based on the moving speed, the steering angle and the area density, and updating a color parameter set; generating a gradient sequence through smooth interpolation; sending a control instruction by using a low-delay protocol and monitoring a delay value; and calculating a comprehensive evaluation value according to the rendering effect, iteratively adjusting the weight if the comprehensive evaluation value exceeds the limit, and regenerating a light control instruction. According to the invention, through the technical means of real-time calculation, dynamic weight adjustment, low-delay transmission, deviation feedback and the like, a light control instruction with smooth gradual change is generated, and accurate control of light color gradual change is realized.

The invention provides a flow velocity identification method and device based on an improved U-Net network model, and belongs to the technical field of flow velocity measurement, and the method comprises the steps: carrying out the sampling of collected river image data, and obtaining the sampling image data; randomly cutting the sampling image data according to a preset size to obtain a cut image; performing color transformation on the cut image according to a preset random probability to obtain an image sequence, the color transformation including at least one of luminosity transformation, contrast transformation, saturation transformation and tone transformation; and inputting the image sequence into an improved U-Net network model to carry out flow velocity identification so as to obtain a flow velocity prediction result of the river channel. According to the flow velocity identification method based on the improved U-Net network model, the time sequence information and the texture information of the river channel video frame are extracted through the improved U-Net network model, and then the corresponding flow velocity value is output, so that the identification precision and the identification efficiency of the flow velocity can be improved.

The invention discloses an SAR image orientation sensitive target detection method based on sub-aperture image saturation characteristics, and belongs to the technical field of synthetic aperture radar image target detection. The azimuth sensitive target detection is realized through the following steps: 1) sub-aperture image generation: dividing an azimuth frequency spectrum into a plurality of sub-bands, and generating a plurality of sub-aperture images through inverse Fourier transform; 2) color synthesis and feature extraction: allocating different hues to each sub-aperture image by using an HSV color space, synthesizing an RGB color image, converting the RGB color image to the HSV color space, and extracting a saturation channel in the RGB color space as an azimuth sensitivity feature map; and 3) target detection: carrying out threshold segmentation on the saturation feature map, and identifying a pixel region with a high saturation value, namely, an orientation sensitive target. According to the method, the color saturation change caused by the scattering difference of the target in different sub-apertures is utilized, effective detection of azimuth sensitive targets such as artificial buildings and vehicles is achieved, and the method has the advantages of being simple in calculation and high in robustness.

The invention relates to a balanced color perception enhancement method for rail transit target detection, and belongs to the technical field of rail transit and computer vision. According to the method, brightness channel adaptive histogram equalization enhancement CLAHE-LC, a multi-segment tone channel mask mechanism MSHCM-M and a three-stage hybrid mechanism non-maximum suppression TSLSH-NMS technology are combined, so that the problems of complex illumination, multi-target shielding, color interference and the like in a rail transit scene are solved, and the accuracy and recall rate of target detection are improved. The method can be seamlessly integrated into an existing deep learning target detection framework, is compatible with multi-label, multi-category and multi-scale features, and remarkably improves the precision and recall rate of target detection in a rail transit scene. Experimental verification shows that the real-time performance is guaranteed, meanwhile, the detection performance is obviously improved compared with a traditional scheme, and the method is suitable for being applied to an actual rail transit safety monitoring and intelligent maintenance system.

A system for the automated evaluation of tobacco leaves, comprising: • a mechanical feeding module (201) configured to feed individual tobacco leaves one after the other; • a controlled imaging chamber (202) equipped with a high-resolution camera and calibrated light sources to capture detailed images of tobacco leaves under standardized lighting conditions; • a color sensor (203) configured to extract colorimetric data from the tobacco leaves, including, but not limited to, RGB values and hue information; • a processing unit (204) comprising one or more processors and memories configured to: i) performing image pre-processing including noise reduction and contrast enhancement, (ii) Execution of segmentation algorithms to isolate the leaf from the background, (iii) Extraction of visual features such as size, shape, texture and color uniformity from the images, (iv) Analyzing the sensor data using statistical or machine learning models to determine quality parameters, (v) Integration of the extracted image features and colorimetric data into an evaluation algorithm based on predefined quality standards, (vi) to assign a quality mark to each tobacco leaf on the basis of the analysis; • an output module (205) configured to display the classification result in real time via an LED screen or a digital interface; • a storage module (206) configured to store classification results for future reference and analysis; • and an Internet of Things (IoT) communication module (207) that enables remote access, monitoring and control of the classification process.

The invention relates to the technical field of image generation, in particular to a content generation method and device based on a user image and a medium, and the method comprises the steps: collecting a user face image and environment features, and extracting face key points, hair style features, expression intensity parameters and posture features of a user in the face image; generating an image interaction state vector in combination with the environment features; presenting a stimulus image sequence containing different tones and colors to the user, collecting facial micro-expressions of the user, establishing a facial expression response curve, and further generating a color physiological response curve; extracting specified characteristic parameters based on the color physiological response curve, and constructing a visual image personality vector in combination with the image interaction state vector; and inputting the personality vector of the visual image into an image style parameter mapping network to obtain an image style control parameter set, processing the three-dimensional reconstruction model of the user face by an image generation and rendering module according to the image style control parameter set, and outputting a static image or a dynamic image.

The invention discloses a self-adaptive low-illumination image enhancement method and device and a storage medium, which are used for stably outputting a high-quality image under various complex conditions. The method comprises the steps of generating an adaptive key parameter set through a preset mapping function based on brightness features; processing the brightness channel through a combined filtering strategy and carrying out adaptive weighted fusion to obtain an irradiation component; converting the brightness channel and the irradiation component to a logarithm domain and performing operation to obtain an initial reflection component; performing linear stretching processing on the initial reflection component to obtain an enhanced reflection component; performing nonlinear gamma correction processing on the irradiation component to obtain a corrected irradiation component; converting the enhanced reflection component back to a linear domain, and performing operation with the corrected irradiation component to obtain an enhanced brightness channel; and recombining the enhanced brightness channel with the original hue channel and saturation channel in the HSV color space, and converting back to the RGB format to obtain an enhanced image.

The invention provides a semantic segmentation and image enhancement method and system for a complex background of a transformer substation, and relates to the technical field of computer vision, and the method comprises the following steps: collecting original image data of the transformer substation, and combining heat map saliency analysis and a spectrum residual error method based on a reciprocal function to obtain a semantic segmentation and image enhancement result; preprocessing the collected original image data to obtain a hue saturation brightness image; performing brightness component extraction on the hue saturation brightness image, and establishing an end-to-end image enhancement model based on an IAT model; performing brightness component enhancement by using the image enhancement model to obtain an enhanced image; and performing diffusion processing on the enhanced image based on a diffusion coefficient function, and performing semantic segmentation processing on the image after diffusion processing by using a semantic segmentation model in combination with a Laplacian operator to obtain a semantic segmentation result. According to the method, image distortion and excessive problems caused by image enhancement are avoided, and the readability of the image is ensured, so that a reliable data basis is provided for subsequent analysis.

An image capturing apparatus for capturing an object and generate an image, displaying the image, performing signal processing for the image, and accepting an input from a user, wherein the image is displayed with a vectorscope superimposed thereon, the vectorscope indicating a distribution of color components included in the image based on a hue and a color density and an index for specifying a range of a skin color region in the image, a display position of the index on the vectorscope is specified based on the hue and the color density values, and a size of the index is specified by the range of the skin color region value, and processing is performed for decreasing a detail level of an image signal of the image having a color component included in the index.

The invention relates to methods, devices, and software for distinguishing hemoglobin (Hb) from myoglobin (Mb) in urine samples using peroxidase-reactive test strips. A time-resolved video of the color change on the test pad is recorded, and hue values are extracted and analyzed over time. Diagnostic differentiation between Hb and Mb is achieved by evaluating the hue-time curve, in particular via endpoint hue values and the ratio of integrated hue values in distinct time intervals. The invention further comprises a standardized optical setup for recording, a software-based evaluation algorithm, and a refillable cartridge for in-situ urine analysis devices containing appropriate test fields. The invention enables combined testing with additional urine parameters and is suitable for medical and non-clinical applications such as sport and home diagnostics.

The embodiment of the invention relates to the field of color matching, and discloses a color matching method and tool for pigment. The method comprises the following steps: acquiring a first brightness value of a target color; based on a black pigment and a white pigment, blending a first formula pigment according to a preset proportion corresponding to the first brightness value; wherein the brightness of the first formula pigment is the first brightness value; blending a second formula pigment based on the basic hue pigment; wherein the brightness of the second formula pigment is the first brightness value, and the hue of the second formula pigment is the hue of the target color; and mixing the first formula pigment and the second formula pigment to obtain the final formula pigment. Therefore, through the scientific color matching method, the color matching accuracy and efficiency of color matching personnel are improved.

The invention discloses a night enhanced imaging and dynamic denoising method for an automobile data recorder, relates to the technical field of vehicle-mounted image processing and computational photography, and is used for solving the problems of insufficient brightness and too strong noise of night videos of the automobile data recorder under a low-illumination condition. Estimating the zero-mean Gaussian noise intensity of a read link and the Poisson noise intensity increasing along with the brightness in the same frame, converting the zero-mean Gaussian noise intensity and the Poisson noise intensity into brightness equivalent parameters, and introducing a unified control function G as the scale reference of all enhancement and noise reduction links; defining a sharing priori H and carrying out unified constraint in dense optical flow, depth time sequence denoising and local tone mapping; and in combination with the motion mask and the background model, performing differential space-time denoising and proportional write-back, and in cooperation with motion blur detection, spatial variation motion kernel and constrained deconvolution and multi-frame detail recharge, obtaining target brightness and outputting an RGB frame after proportional write-back.

The invention belongs to an image processing technology, and particularly relates to an underwater image enhancement method and system based on combination of multiple convolution models and an attention mechanism, and the technology is formed by combination of a DIAM model and a Shallow-UWnet improved model. The DIAM model is mainly used for carrying out color restoration in a mode of complementing hue weakening and non-uniformity of the underwater image, and can solve the problems of underwater noise, atomization and the like. According to the Shell-UWnet improved model, on the basis of Shell-UWnet, a highest pooling layer and an Inc multi-convolution module are added, the highest pooling layer effectively reduces network parameters and calculation overhead through dimension reduction operation, the operation efficiency is improved, the key feature extraction capacity is enhanced, detail loss is avoided, Inc can extract features from different scales, and the algorithm is simple and convenient to operate. And the feature capture capability of the model is enhanced.

The invention discloses a camouflage confrontation coating generation and physical enhancement method, which comprises the following steps: generating a uniform Thiessen polygon based on an RGB-D image, and calculating control point distribution regularization loss; fusing the multi-scale RGB and depth features; the color distribution probability is determined by combining the pixel point distance and depth information, and preliminary coating is generated; the camouflage concealment is optimized through saliency detection and environment dominant hue fusion; spatial domain affine transformation and RGB-CMYK transformation are introduced to enhance physical robustness, and visual consistency loss is calculated; and optimizing a comprehensive loss function by combining control point distribution regularization loss, multi-scale fusion saliency loss, visual consistency loss and target detection confrontation loss, and iteratively generating final coating. According to the method, the geometric camouflage precision is improved, the visual saliency of physical coating is reduced, the environmental adaptability and the anti-robustness are enhanced, and the method is suitable for hiding and anti-attack of a complex surface target.

The invention provides a DFKD new framework combining an intra-class and intra-class decoupling constraint mechanism and a cross-class attention diffusion migration technology, and aims to simultaneously realize high-quality style diversification and semantic content consistency of a synthetic sample under the condition of completely no external data. Specifically, the main purpose of the invention is as follows: by designing intra-class-inter-class style unwrapping constraint loss, a generator is prompted to realize style diversity enhancement of different style samples in the same class, and meanwhile, the distinction degree of style features among different classes is ensured; according to the method, cross-layer attention features in a teacher model are utilized to drive a diffusion model to carry out style migration on multiple scales, and multi-dimensional style changes such as hue, texture and illumination can be flexibly captured without any real image reference. Results show that the method reaches the current advanced level in the field of data set distillation, and shows excellent convergence speed in various data set distillation methods.

The invention provides a deep learning-based automatic toning system and method for a record. The method comprises the following steps of: determining a toning constraint condition of the record in visual constraints according to style similarity of shallow visual features between adjacent recorded image frames in the record and brightness difference between the adjacent recorded image frames; performing fuzzy correlation on the scene semantic elements of the narrative segments in the record through the cross correlation among the scene semantic features to obtain the element membership of the scene semantics in the narrative segments, and further determining the scene attention of different narrative segments in the record according to all the element membership; and according to the scene attention and the hue constraint conditions of the different narrative segments, determining the chroma reconstruction cost of the record in the toning process, and carrying out automatic toning processing on the record based on the chroma reconstruction cost. According to the scheme, the visual features and the semantic elements can be fused in the automatic toning process of the record, and toning coordination with narrative logic as the guide is achieved.

The invention discloses an atmosphere lamp control method and device based on visual perception and an automobile, and the method comprises the steps: partitioning a display screen, obtaining a display partition matrix, and enabling partition elements or partition element combinations in the display partition matrix to correspond to corresponding lamp strips; performing multi-dimensional saliency scoring on the candidate color of each partition by using a K-means clustering module, a vision priority detection module or a spatio-temporal context weighting module according to the preprocessed display image, and determining a typical color according to the saliency comprehensive score of each candidate color; and respectively mapping the typical color of each display subarea to the corresponding lamp strip. Visual intelligent perception is achieved, the intelligent degree of atmosphere lamp control in the vehicle is improved, and the problems that in an existing automobile atmosphere lamp control technology, screen partition is simple, color picking is not intelligent, simple screen average color picking or partition color picking is mostly adopted, dominant hues and key details perceived by human eyes cannot be accurately reflected, and the color picking efficiency is poor are solved. And the light color and the picture content are separated.

To provide an optical information learning generation apparatus which, in generation of a two-dimensional image from optional viewpoints of a three-dimensional scene using different cameras, contributes making it possible to generate a two-dimensional image with same color regardless of difference of colors resulting from used cameras.SOLUTION: An optical information learning generation apparatus according to the present invention has an optical information learning estimator including an optical learning estimation unit for learning optical information by using, as a teacher data, a two-dimensional image of a three-dimensional scene acquired by a sensor, a sensor identifier input unit for receiving a sensor identifier for use in identifying the sensor, a spectral characteristic conversion unit, and a spectral characteristic setting unit for setting spectral characteristics to the spectral characteristic conversion unit. The spectral characteristic setting unit sets a learning spectral characteristic corresponding to the sensor identifier to the spectral characteristic conversion unit. The optical learning estimation unit outputs optical information to the succeeding part. The spectral characteristic conversion unit, in accordance with the learning spectral characteristics, output a color value based on the optical information.SELECTED DRAWING: Figure 1

The embodiment of the invention provides a sewage draining exit state identification method and system, and can solve the problems that the traditional sewage draining exit monitoring mainly depends on manual inspection, the efficiency is low, the coverage is limited, and a hidden sewage draining exit is difficult to find in time. The method comprises the following steps: acquiring image data of a to-be-identified area through an unmanned aerial vehicle; image data of the to-be-recognized area is input into a sewage draining exit state recognition model, the sewage draining exit state recognition model is obtained through training based on sewage draining exit image data under different marked sewage draining states, different shooting angles and environment conditions, and a model input channel is internally provided with a water color feature branch; the RGB extraction module is used for extracting RGB values of a target area and a surrounding water body in the image, and converting the RGB values into hue H and saturation S in an HSV color space through nonlinear transformation; and based on the sewage draining exit state identification model, determining a sewage draining exit position, a sewage draining state and a confidence coefficient corresponding to the sewage draining state in a to-be-identified area.

Some embodiments of the present application show a display device and an image display method, the method comprising: identifying an image content of a to-be-displayed image, and obtaining an image parameter based on the image content, the image parameter being used for representing a hue or contrast of the to-be-displayed image; calculating brightness adjustment values of a plurality of color light-emitting units in a light-emitting unit group corresponding to pixels in the to-be-displayed image according to the image parameters; and controlling a backlight controller to adjust the brightness of a plurality of color light-emitting units in the light-emitting unit groups corresponding to the pixels according to the brightness adjustment values. According to the embodiment of the invention, the color backlight source is selected, and the brightness of the different color light-emitting units corresponding to each pixel in the to-be-displayed image is flexibly and independently adjusted according to the image parameters, so that the contrast ratio of the image is enhanced, and more accurate color restoration is realized.

A device for image processing includes processing circuitry configured to perform black level remapping. To perform black level remapping, the processing circuitry is configured to: determine luminance values for pixels of a first image, wherein the pixels of the first image are based on color values of samples in each color frame of a plurality of color frames of the first image, wherein each color frame is associated with a different color; determine a first threshold based on the luminance values; determine a mapping curve based on the first threshold and a second threshold; apply the mapping curve to color values of a set of the samples in each color frame to generate black level remapped samples for each color frame, the set of the samples having a color value between the first threshold and the second threshold; and generate a second image based on the black level remapped samples.

The invention relates to a gradient migration and low light enhancement collaborative infrared and visible light image fusion method. The method comprises the following steps: preprocessing an infrared image and a visible light image, converting the preprocessed visible light image into an HSV color space, and extracting a brightness component; performing low-light enhancement and high-light suppression on the brightness component by using a complementary logarithmic gamma function to obtain an enhanced brightness component; carrying out sharpening processing on the enhanced brightness component by adopting a brightness protection dynamic histogram equalization technology to obtain an optimized brightness component; and based on a gradient transfer fusion algorithm, fusing the component with the infrared image to obtain a fused brightness component, merging the fused brightness component with a hue component and a saturation component of the visible light image, and converting into an RGB color space to obtain a fused image. According to the method, through collaborative optimization of gradient migration and low light enhancement, the quality of the fused image under the condition of low light at night is remarkably improved, and the target saliency and the detail retention capability are enhanced.

The invention provides an image processing method and device, electronic equipment and a storage medium, and the method comprises the steps: adjusting a preset adjustment coefficient matrix corresponding to each image quality parameter of an original image in response to an image quality parameter adjustment operation, and generating a target adjustment coefficient matrix of the original image; the image quality parameters comprise at least one of brightness, saturation, contrast, hue and color temperature; converting the original image into a first color space to obtain first image data; based on each target adjustment coefficient matrix, performing image quality adjustment on the first image data to obtain target image data; and sending the target image data to the field programmable gate array chip, and processing the target image data through the field programmable gate array chip to generate a target image of the original image. Based on the embodiment, hardware resources of a chip are saved to a great extent in the color space conversion process, the color space conversion efficiency is improved, and the image output efficiency of the image output equipment is improved.

The invention discloses a low-illumination image enhancement method based on semantic perception and differentiable tone mapping, and belongs to the field of computer vision and image processing. According to the method, semantic information of each region in an image is extracted through a semantic segmentation network, different importance weights are distributed for different semantic regions, and then a differentiable tone mapping curve generation module is driven to dynamically adjust brightness and contrast enhancement parameters of each region. According to the system, an end-to-end trainable architecture is adopted, semantic perception loss, physical consistency loss and image quality loss are jointly optimized, and finally content self-adaptive enhancement of low-illumination images is achieved. Compared with the prior art, the problems of semantic information loss, overexposure or insufficient enhancement and the like in a traditional method are effectively solved. By introducing semantic perception driven by deep learning and differentiable tone mapping, the detail and color consistency of a semantic region can be more accurately kept in the image enhancement process. In the aspect of performance evaluation, the peak signal-to-noise ratio (PSNR) of the method is improved by more than 3.2 dB, the average intersection-to-union ratio (mIOU) of semantic segmentation is improved by more than 8.5%, and the method is remarkably superior to an existing mainstream image enhancement method.

Provided are a color correction matrix optimization method, an electronic device, and a medium. The method includes performing correction processing on a color source matrix of an acquired color chart image based on a color correction matrix to obtain an output color matrix; determining a global hue error and a saturation constraint regularization term based on Lab coordinates of the output color matrix and Lab coordinates of a target color matrix of a target color chart image and constructing a loss function; and updating the color correction matrix based on a value of the loss function until the value of the loss function converges to obtain an optimal color correction matrix to perform color correction on a to-be-processed image based on the optimal color correction matrix.

The invention discloses a mulberry twig clamping and positioning method and system, belongs to the technical field of intelligent mulberry leaf picking equipment, and aims to solve the technical problems that an existing RGB camera is easily interfered by illumination, and picking damage is caused due to the fact that a demarcation point of a tender branch and an old branch is difficult to accurately recognize. According to the method and system, 500-600 nm visible light and 750-900 nm near-infrared band images are synchronously collected through a multispectral imaging device, HSV color space parameters are extracted, and a color gradient model is established; by analyzing the gradient change rate of the hue value H and the saturation degree S and combining the curvature change rate and the near-infrared reflectivity difference value, the transition boundary of the tender and old branches is accurately judged, and the three-dimensional coordinates of the clamping datum point are generated by dynamically shifting 5-15 mm towards the side of the tender branches. According to the scheme, the limitation of single hue detection is broken through, the xylem recognition capability is enhanced by utilizing multispectral information, the clamping accuracy is improved, the method is suitable for automatic operation of a top-remaining picking mode in large-scale silkworm and mulberry planting, the picking efficiency can be effectively improved, and branch damage can be reduced.

The invention provides an image defogging method using color attenuation prior, which comprises the step of preparing a training set and a test set of a RESIDE data set, and comprises the following specific steps of: performing HSVC-Net network training by using an image in the training set to generate a training model; storing the training model in a local folder, testing the effect of the training model by using the images in the test set, and if the training model is satisfied, storing the training model as a satisfied training model; and storing the satisfactory training model in a local folder, and testing a foggy image by using the satisfactory training model. The invention provides a novel image defogging network of an HSVC-Net. Wherein the hue correction module corrects the hue periodic characteristics by using Fourier coding. And the double-domain refinement module realizes HSV and RGB feature interaction through deformable convolution and cross-modal attention. The sparse attention module fuses global sparse attention and local detail supplementation, dynamically captures multi-scale haze deformation and inhibits noise, and realizes high-frequency information retention and color restoration.