605 results about "Remote sensing image processing" patented technology

The invention relates to the technical field of remote sensing image processing, and discloses a remote sensing image adaptive identification method and system for territorial space planning, and the method comprises the steps: obtaining a multi-source remote sensing image data set of a research region, feature extraction, cloud detection, quality evaluation and adaptive preprocessing are carried out; carrying out prototype network coding, calculating a category prototype and probability, and supporting fine tuning of a set; carrying out multi-scale cavity convolution and category scale attention fusion; evaluating the adaptability score of the comprehensive fusion feature map set, and carrying out weighted fusion, classification and normalization; change detection is carried out, stable and change regions are segmented, and time sequence context features are extracted and constrained optimization is carried out; entropy is fused, a boundary is decided, uncertainty is estimated, and weighted fusion is carried out according to a change area; conditional random field optimization, confidence level grading and connected domain identification are carried out; the automation level, the adaptive capacity and the recognition reliability of remote sensing monitoring of territorial space planning are improved.

The invention relates to the technical field of artificial intelligence and remote sensing image processing, and particularly provides a hyperspectral image and laser radar data classification method based on a dynamic fusion network. The method comprises the following steps: preprocessing acquired multi-modal data, and constructing multi-scale input; a dual-scale local attention module is designed, and context information of different scales is fused in a self-adaptive weighted mode through gating soft pooling; a dynamic down-sampling feature enhancement module is designed, the down-sampling rate is dynamically adjusted according to the complexity of the feature map, and deep multi-scale interaction is carried out based on a Mama backbone; constructing a directional interactive attention module, extracting features in horizontal, vertical and diagonal directions through directional gating convolution, and capturing an anisotropic structure of a linear ground feature; through the design of a double-path classifier, fusing shallow space details and deep semantic information; and the model is trained, optimized and reasoned to obtain data classification, and the method improves the classification precision and the calculation efficiency.

The invention discloses an agricultural land utilization monitoring management system and method based on remote sensing, and belongs to the technical field of remote sensing image processing. Multi-temporal remote sensing images are acquired, and a land surface energy fluctuation spectrogram is constructed; extracting disturbance characteristics through small-scale grid slices to form a multi-dimensional disturbance characteristic tensor; identifying an abnormal evolution region by using a sparse volume accumulation algorithm, and outputting a preliminary screening identification graph; inputting the region with the continuous evolution characteristic into a time sequence attention mechanism inversion network, estimating a crop growth state trajectory, matching an agricultural planting mode library, and generating a candidate land utilization behavior probability distribution diagram; in combination with regional consistency optimization, a historical planting period and meteorological disturbance data, calculating a purpose change confidence score, and outputting early warning information and a monitoring report; the agricultural land dynamic change identification precision and management capability can be effectively improved.

The invention discloses a remote sensing image space-time fusion method and device based on a selective state space model, and belongs to the technical field of remote sensing image processing and computer vision crossing. The method comprises the following steps: acquiring high-resolution and low-resolution image input, and extracting multi-scale features through a multi-layer encoder; capturing an anisotropic space structure in the remote sensing image by using a four-way two-dimensional selective scanning mechanism; designing a state space fusion module, decoupling and cooperatively processing space details and time dynamic information through a space and time sequence selective scanning fusion sub-module, and performing feature fusion by adopting adaptive gating parameters; and finally, reconstructing a high-resolution image through a symmetric decoder, and carrying out model optimization by adopting a composite loss function. On the premise of ensuring the linear calculation complexity, the spatial detail fidelity, the time continuity and the overall efficiency of the fused image are remarkably improved, and the method is suitable for large-scale remote sensing data processing.

The invention relates to the field of remote sensing image processing, and provides a remote sensing image terrain radiation correction method, device and equipment, and the method comprises the steps: obtaining a to-be-corrected remote sensing image, calculating the terrain parameter of each pixel, dividing the to-be-corrected remote sensing image into a shadow region and a non-shadow region, and determining the direction weighting function of each pixel; calculating a topographic modulation spherical harmonic primary function corresponding to each pixel; training samples are selected in the shadow area and the non-shadow area respectively, a linear model is constructed, and spherical harmonic illumination coefficients representing overall three-dimensional illumination distribution of the shadow area and the non-shadow area are solved; and based on the topographic modulation spherical harmonic primary function, combining the spherical harmonic illumination coefficient corresponding to the area to which each pixel belongs, reconstructing the average incident radiance of each pixel, and then performing inversion to obtain the real reflectivity of the earth surface. According to the method, the problem of inaccurate radiation distortion correction of the remote sensing image under the complex terrain in the prior art is solved, and accurate inversion of the real reflectivity of the earth surface is realized.

The invention relates to a rice identification method based on optical and SAR image fusion, and belongs to the technical field of remote sensing image processing. The method comprises the following steps: synchronously acquiring SAR and optical time sequence data, respectively preprocessing the SAR and optical time sequence data, and extracting characteristic parameters; the method comprises the following steps: constructing a double-flow Vision Transform model, wherein the double-flow Vision Transform model comprises a double-branch feature extraction module, a space-time Transform aggregation module, a gating feature fusion module and a decoder; the double-branch feature extraction module extracts SAR and optical features according to the SAR and optical feature parameters; inputting the extracted SAR features and optical features into a space-time Transform aggregation module, and carrying out time sequence dependence modeling through a self-attention mechanism; the gating feature fusion module performs weighted fusion on the optical and SAR features; and the decoder generates a rice distribution probability graph according to a weighted fusion result. According to the method, the rice identification precision under the multi-temporal and multi-environment conditions can be improved.

The invention relates to the technical field of remote sensing image processing, and provides a remote sensing multi-modal target detection method and device for gated cross attention fusion, and the method comprises the steps: obtaining a multi-modal remote sensing image pair; and inputting the multi-modal remote sensing image pair into the target detection network model, and outputting a target detection result. According to the remote sensing multi-mode target detection method based on gating cross attention fusion, firstly, the edge information of each mode is explicitly enhanced, then, the information interaction and reconstruction between the two modes are realized by using a bidirectional cross attention mechanism, a key region can be adaptively highlighted, redundant or noise characteristics can be inhibited, and the detection accuracy is improved. The reconstructed visible light feature map and invisible light feature map not only maintain respective modal advantages but also have complementarity, dynamic weighted fusion is performed on different modal features through a gating weighted fusion mechanism, target discrimination features in the multi-modal fusion feature map are highlighted, accurate and stable target detection is realized, and the target detection accuracy is improved. And the advantages of the multi-modal image can be fully played in a complex scene.

The invention relates to the technical field of computer vision and remote sensing image processing, in particular to a super-resolution remote sensing image target detection method based on multi-modal fusion, which is realized on the basis of a target detection network obtained by improving a YOLOv5 network. The improvement comprises the following steps: introducing a dynamic cross calibration module and a lightweight super-resolution auxiliary branch behind an input layer; replacing part of common convolution in the backbone network with space-to-depth convolution; the method comprises the following steps: inputting a visible light image and an infrared image into a dynamic cross calibration module, carrying out feature extraction and fusion on the visible light image and the infrared image through the dynamic cross calibration module, and outputting fusion features; inputting the visible light image and the infrared image into a lightweight super-resolution auxiliary branch, and performing super-resolution reconstruction through the branch to generate a high-resolution feature; the problems that an existing remote sensing image target detection method is low in accuracy and prone to loss of detail features when facing small targets, low resolution and complex backgrounds are solved.

The invention relates to the technical field of remote sensing image processing, and particularly discloses a low-light remote sensing image restoration method and system based on double-frequency domain processing, and the method comprises the steps: constructing an image restoration network which comprises a coding module, an intermediate enhancement module and a decoding module which are connected in sequence, the coding module and the decoding module are in jump connection; wherein the coding module, the intermediate enhancement module and the decoding module are each internally provided with a double-frequency-domain attention module, and each double-frequency-domain attention module comprises a Fourier attention sub-module used for global frequency domain feature modeling and a wavelet attention sub-module used for multi-scale detail feature extraction; by introducing the Fourier transform frequency domain processing technology, the global frequency characteristic analysis capability is provided, and efficient global modeling is realized. Secondly, introducing a wavelet decomposition frequency domain processing technology, decomposing the image into sub-bands with different scales and frequencies, and effectively separating a clear image and a degenerated component;

The invention belongs to the technical field of geographic information technology and remote sensing image processing and service, and discloses a slice-free image service system and method based on dynamic projection and real-time mosaic, which abandons the traditional pre-slicing mode, directly provides online service based on original image data, and improves the service efficiency. The problem of redundant storage caused by pre-generation and storage of massive tiles is fundamentally eliminated, and storage space is saved by up to 90%. Meanwhile, when the original data is updated, the system does not need to carry out a time-consuming re-slicing process, real-time updating and publishing of services can be realized, and the core pain points of long data updating period and high delay in the traditional technology are thoroughly solved. Through the integrated dynamic projection engine and the real-time mosaic module, the on-demand service request of the client for any coordinate system, any spatial range and any resolution can be responded.

The invention discloses a multi-source remote sensing image classification method fusing a frequency domain attention mechanism and a cross-modal Transform, and relates to the technical field of remote sensing image processing. The method comprises the following steps: performing frequency domain enhancement on a hyperspectral image through a frequency domain attention mechanism; semantic enhancement is carried out on the laser radar image through depth separable convolution operation and a multi-head self-attention mechanism; respectively extracting deep semantic features corresponding to the two modal enhancement features through a Transform encoder; respectively weighting the deep semantic features of the two modals through a channel attention branch and a space attention branch, and performing multi-level feature fusion by using a learnable weight to obtain a fusion output feature; and aggregating and fusing the global semantic vectors of the output features in two modes of global average pooling and attention pooling, splicing the two global semantic vectors, and mapping the spliced global semantic vectors to a category space to obtain a ground feature classification result of the multi-source remote sensing image. According to the method, the classification precision under the conditions of complex city scenes and small samples is improved.

The invention relates to a remote sensing visual language large model training method and device based on unified reinforcement learning, and belongs to the technical field of artificial intelligence and remote sensing image processing, and the method comprises the steps: carrying out the preprocessing of an input remote sensing image and a text instruction, and extracting visual features and text features; performing modal alignment, and inputting the modal alignment result into a pre-trained large language model for supervised instruction fine adjustment to obtain a basic model; constructing a multi-dimensional deterministic unified reward module based on a truth value; and performing enhanced fine tuning on the basic model, calculating a reward value output in the group by using a unified reward module, and updating model parameters based on relative advantages to obtain an optimized remote sensing visual language large model. According to the method, a deterministic reward module is adopted, a value network is abandoned through group relative strategy optimization, the calculation cost is reduced, task-level indexes are directly optimized through a multi-dimensional unified reward function, and the precision and output normalization of remote sensing image interpretation are improved.

The invention discloses an SAR directed target detection method based on multi-scale context sensing, and belongs to the technical field of remote sensing image processing and computer vision. According to the method, firstly, through a scattering characteristic guided multi-scale dynamic characteristic enhancement network, a multi-branch structure and a dynamic fusion mechanism are utilized to enhance characteristic expressions of targets of different scales; and secondly, designing a task-adaptive regional context sensing module, fusing local details and semantic contexts, and improving the target discrimination capability under a complex background. And a decoupling type progressive refining detection head is adopted to predict a target category, a directed bounding box and a direction angle, and the positioning precision is optimized through progressive regression. Semantic consistency loss is introduced in the training stage, and end-to-end optimization is achieved in combination with a multi-task joint strategy. According to the method, the problems of large target scale difference, strong background interference, changeable directions and the like in the SAR image are effectively solved, and the accuracy and robustness of directed target detection are remarkably improved.

The invention relates to the technical field of remote sensing image processing, in particular to an image super-resolution system and method based on high and low frequency separation perception Mama, and the method comprises the steps: firstly carrying out the shallow convolution feature extraction of a low-resolution image; then entering a plurality of frequency sensing Mama groups, performing frequency separation and enhancement on each group through a high and low frequency feature adaptive enhancement module, and performing depth feature transformation through a plurality of frequency sensing Mama blocks; the extracted depth features are refined through a global channel-space attention module, and finally a high-resolution image is reconstructed through up-sampling. Through organic combination of the modules, the defects of insufficient frequency perception, low global modeling efficiency, insufficient feature optimization and the like are effectively overcome, and high-quality collaborative reconstruction of remote sensing image structures and textures is realized.

The invention belongs to the technical field of remote sensing image processing, particularly relates to an urban remote sensing image segmentation method and system based on bidirectional coordinate attention and multi-scale adaptive feature fusion, and provides a remote sensing image semantic segmentation neural network architecture taking an attention re-calibration module as a decoder core. Wherein the encoder path gradually extracts multi-scale feature representation through cascaded residual convolution blocks and down-sampling operation to form a feature pyramid of which the spatial resolution is reduced step by step and semantic information is enhanced step by step; the decoder path gradually recovers the spatial resolution through cascaded up-sampling and feature refining operations to generate a precise segmentation mask; a space-channel dual attention re-calibration module oriented to a decoding stage is provided, through explicit coding of space coordinate direction information and combination of global channel dependence modeling, features beneficial to semantic discrimination are adaptively enhanced in the feature fusion and resolution recovery process, and therefore segmentation precision and consistency are improved.

The invention discloses a segmented multi-source shoreline extraction and calibration method and system based on physical prior, and particularly relates to the technical field of remote sensing image processing and computer vision, a consistency evaluation and reliability measurement mechanism of a multi-source surface water product is integrated on a data level, and a high-quality training sample set is constructed by using water persistence characteristics; on the model level, a feature extraction framework fusing physical prior guidance and a geometric self-adaptive operator is researched and developed, a multi-dimensional physical constraint and noise tolerance loss function is constructed, and pixel-level accurate recognition and topological rigorous quality closed-loop calibration of a shoreline under a large-range remote sensing image are achieved.

The invention relates to the technical field of crossing of remote sensing image processing and computer vision, and discloses a remote sensing image change detection method which comprises the following steps: obtaining a first time phase bottom layer pixel feature and a second time phase bottom layer pixel feature according to a first time phase remote sensing image and a second time phase remote sensing image; according to the first time phase bottom layer pixel features and the second time phase bottom layer pixel features, enhanced visual features are obtained; according to the first pixel-text similarity score plot and the second pixel-text similarity score plot, semantic guidance features are obtained; according to the enhanced visual features and semantic guidance features, obtaining fusion features of multiple levels; pixel-level binary classification is carried out on the fusion features of the multiple levels, and a change detection binary image is output. According to the invention, the precision and accuracy of remote sensing image change detection are improved.

The invention discloses an unmanned aerial vehicle remote sensing vegetation recognition system and method based on multi-scale segmentation, and particularly relates to the technical field of remote sensing image processing. Constructing a differentiable physical illumination model to generate a theoretical incident light intensity distribution diagram; channel superposition is carried out on the illumination distribution map and an original image, and fusion features are extracted through a multi-scale convolutional neural network; further, a feature decoupling mechanism guided by a landform and vegetation coupling coefficient is utilized to adaptively suppress illumination and terrain interference; and finally, realizing accurate division of vegetation categories through a multi-path fusion decoder. Through deep fusion of a physical mechanism and deep learning, the problems of low vegetation identification precision and poor model generalization ability caused by illumination variation in a complex terrain environment are solved.

The invention discloses an adaptive feature alignment crop identification method based on intelligent remote sensing interpretation, and the method comprises the steps: taking DINOv3 as a backbone network, and optimizing the backbone network through a feature normalization adaptation layer and a resolution adaptive module; a time sequence feature alignment module TAFA is adopted to carry out cross-time alignment on crop features of remote sensing images in different growth periods; a geographic context gating attention module GCAI is adopted, and cross-geographic region adaptive feature fusion is realized based on multi-scale geographic context coding and a CNN-Transform bidirectional interaction mechanism; a crop semantic contrast loss function CSCL is adopted to perform intra-class compactness and inter-class separation degree contrast optimization on features of crops of the same class. The method realizes cross-growth-period stability and cross-region adaptation under a pure supervised training normal form, can distinguish remote sensing crop high-precision semantic segmentation of similar crops, covers agricultural remote sensing image processing scenes of different geographical landforms and whole growth periods of crops, and can be applied to the fields of modern agricultural fine management and the like.

The invention relates to the technical field of remote sensing image processing, in particular to a remote sensing image semantic segmentation method fusing boundary guidance and a semantic compensation mechanism. Semantic segmentation is carried out by constructing a boundary guidance semantic compensation network comprising an encoder based on EfficientNet-B3, a decoder based on Transform, a cross-layer semantic compensation module and an auxiliary boundary supervision module, after an original remote sensing image is input, hierarchical features with strong semantic information and different spatial resolutions are extracted through the encoder, and the hierarchical features are extracted through the boundary guidance semantic compensation network. The method comprises the following steps: firstly, generating a fusion feature with high-level semantics and fine space details by using a cross-layer semantic compensation module, then integrating a global context by using the powerful modeling capability of Transform, gradually reconstructing high-resolution semantic mapping through the fusion with compensated features, and finally outputting a segmentation result. In the training process of the boundary guidance semantic compensation network, the boundary detail loss can be reduced through an auxiliary boundary supervision module, and the segmentation precision and robustness are improved.

The invention relates to the technical field of remote sensing image processing, in particular to a raft laver culture area end-to-end extraction method based on frequency dynamic filtering, and the method comprises the steps: obtaining an original remote sensing image of a raft laver culture area, carrying out the preprocessing, constructing a mask data set for high-level semantic supervision, and carrying out the extraction of the original remote sensing image; constructing a hierarchical supervision neural network model, and performing model training on the constructed hierarchical supervision neural network model, including constructing a composite loss function composed of mask loss, gravitational field loss, vertex classification loss and vertex offset loss; and performing optimization processing on a prediction result output by the hierarchical supervision neural network model, and generating a breeding area contour vector result based on the optimized binary mask and vertex coordinates. The method solves the problem that the extraction precision is suddenly reduced in a scene of relatively strong spatial spectrum characteristic heterogeneity of the breeding area in a traditional method, and improves the extraction accuracy of the breeding area. And reliable technical support is provided for large-area and high-precision monitoring of an offshore raft type laver culture area.

The invention belongs to the technical field of remote sensing image processing and target inversion, and particularly relates to a hyperspectral image aerial aircraft height and spectral reflectivity inversion method based on wake clouds and shadows of the wake clouds. In order to solve the technical problems of low aerial aircraft height inversion precision and inaccurate spectral reflectivity inversion in the prior art, firstly, wake cloud and shadow detection, matching and geometric modeling are performed on a hyperspectral image, and the aircraft flight height is inverted by using a stable geometric relationship between the wake cloud and the shadow; high-precision height extraction is realized; and then combining flight height information to establish a radiation transmission model for coupling the aerial aircraft and the background environment, and inverting the spectral reflectivity of the model. Through integrated processing of wake cloud detection, shadow matching, height inversion and spectral reflectivity inversion, the accuracy and applicability of aerial aircraft characteristic inversion are significantly improved, and an effective technical support is provided for application of hyperspectral remote sensing in aerial aircraft monitoring and identification.

The invention discloses an unmanned aerial vehicle remote sensing image change detection method and system, and relates to the technical field of remote sensing image processing. The method comprises the following steps: acquiring dual-time-phase remote sensing images collected by an unmanned aerial vehicle in the same designated area at different time periods; inputting the remote sensing image into a pre-trained remote sensing image change detection model, and outputting a change detection result graph; wherein the remote sensing image change detection model is configured to extract RGB features and depth features of two time-phase remote sensing images respectively; performing cross-modal fusion on the RGB features and the depth features to obtain enhanced single-temporal features corresponding to each temporal phase; and performing cross-time interaction on the enhanced single-time-phase features of the two time phases to identify and output a change region. According to the method, the utilization of depth information in an unmanned aerial vehicle remote sensing image feature extraction process and depth interaction and recognition between dual-tense features can be enhanced, the influence of false change noise is reduced, and the prediction performance of the model is improved.

The invention relates to a multi-agent-driven remote sensing image intelligent processing algorithm autonomous training verification system. The system comprises a task input and large model agent planning module, a data screening and preprocessing module, an algorithm model screening and super parameter configuration module, a model training and testing module and a comparative analysis and report generation module. According to the method, the whole process of data preparation, algorithm matching, parameter configuration, parallel training, test verification and report generation can be intelligently completed through multi-agent cooperative work, manual intervention is remarkably reduced, the model research and development efficiency and quality are improved, and it is ensured that the process is traceable and the result is reproducible.

The invention belongs to the technical field of remote sensing image processing and computer vision, discloses a semi-supervised ground feature classification optimization method for an urban remote sensing scene in a complex environment, and solves the problems of strong dependence on a large amount of annotation data, incomplete model feature extraction and low calculation efficiency in the prior art. According to the method, firstly, through a cross-view weak-to-strong consistency training framework, a weak enhancement view is utilized to generate a high-confidence pseudo tag, and through applying consistency constraint to a strong enhancement view, a self-optimization training cycle is constructed, so that the dependence on pixel-level annotation data is remarkably reduced. Secondly, a mixed CNN-Mama double-flow framework is adopted, and local texture features are extracted through a lightweight residual error convolution module; according to the method, the classification precision, robustness and generalization ability of the model in a complex city scene are effectively improved, meanwhile, the high efficiency of the method is ensured through lightweight design, and the method is suitable for actual application scenes such as disaster loss assessment and city planning.

The invention discloses a building detection system and method cooperatively driven by remote sensing data and an AI model, particularly relates to the field of remote sensing image processing, and is used for solving the problem of false report and missing report of building change detection caused by the influence of registration errors and imaging style differences on multi-temporal remote sensing images. The method is used for solving the problem of false detection and missing detection of building change caused by the influence of registration errors and imaging style differences on two-time-phase remote sensing images. A registration image pair is generated through a first time-phase remote sensing image and a second time-phase remote sensing image, and a multi-branch prior coding segmentation network outputs a first time-phase building mask, a second time-phase building mask stable region mask building boundary diagram and a building fragment library. The self-inverse style mapping network generates a same-style image pair and constructs a synthetic image pair and a synthetic change label graph, the synthetic samples are credibly judged to screen out uncredible samples, the interactive attention change detection network is trained to obtain a change detection model, a credible change graph is generated through reasoning, and a building fragment library is updated.

The invention belongs to the technical field of remote sensing image processing and deep learning, and particularly relates to a remote sensing erosion gully segmentation method combining multi-scale features and an attention mechanism. Comprising the following steps of 1, data preparation and data preprocessing; 2, constructing and enhancing a data set; step 3, model construction and strategy training; and 4, performing comparison experiment and result evaluation. According to the method, the segmentation precision of the extremely long and thin erosion gully target can be remarkably improved; the topological connectivity and the morphological fidelity of the linear landform are enhanced; pixel-level accurate space positioning and noise suppression are realized; the method has extremely high slant feature capture and multi-scale adaptability; and the stability and weak signal perceptibility of the training process are improved.

The invention belongs to the technical field of remote sensing image processing agricultural information, and particularly relates to a DS-xNet-based multi-source remote sensing time sequence data cultivated land utilization current situation extraction method. According to the method, the cultivated land and the water body can still be reliably recognized under the cloud or fog condition by combining Sentinel-1 and Sentinel-2, missing detection and misjudgment caused by cloud shielding are reduced, sNET captures short-term dynamic conditions, and mNET matrix memory enhances long-term dependence modeling, so that the recognition precision of seasonal and interannual changes is improved, supervision is applied to a branch and fusion layer, and the recognition accuracy of the cultivated land and the water body is improved. According to the method, each branch can learn independent discrimination capability and can cooperatively improve fusion output, over-fitting is reduced, generalization capability is improved, training convergence is accelerated, channel splicing retains discrimination information of each modal in a fusion stage, modal information conflict or loss caused by direct early fusion is avoided, classification accuracy is improved, and classification efficiency is improved. Various optical and radar indexes are combined and used to enhance the distinguishing capability of paddy fields, irrigation areas, different crops and non-cultivated areas.

The invention relates to the technical field of remote sensing image processing, in particular to a few-sample hyperspectral remote sensing image classification method and system based on multi-path evolution. The method comprises the steps of generating a spectral attention weight by utilizing spectral compression reconstruction according to acquired multi-scale deep feature representation, and performing three-dimensional convolution evolution on weighted features based on adaptive multi-path feature evolution to obtain information fusion features; performing weighted fusion on the information fusion features of the three paths through a dynamic gating mechanism; and obtaining a hyperspectral image classification result according to the fused features. According to the invention, self-adaptive evolution and cross-path selective fusion of multi-scale features are realized; the deep features subjected to multi-level feature coding are used for final classification reasoning, so that high-precision hyperspectral image classification is realized under the condition of extremely few samples.

The invention discloses an image defogging method and system based on multi-scale residual attention, and the method specifically comprises the steps: inputting a pre-processed foggy image into a network model based on a U-Net architecture, extracting multi-scale features through an encoder, carrying out the processing of a feature refining module, and generating a fogless image through a decoder, and outputting the fogless image. According to the invention, a multi-scale residual block is designed, convolution kernels of different sizes are used to capture features of different scales, and a residual structure and a mixed attention mechanism are introduced to enhance the feature extraction capability. Meanwhile, a feature refining module is introduced to enhance the image detail recovery capability, and smooth L1 loss and perception loss functions are used to optimize the network performance. The method aims at effectively removing the fog in the image and improving the image quality, and is particularly suitable for the fields of intelligent driving, remote monitoring, remote sensing image processing and the like. The method is obviously superior to the prior art in qualitative and quantitative evaluation, fog can be effectively removed, and image detail textures can be recovered.