203 results about "Fusion image" patented technology

The invention discloses a medical image intelligent processing method and system based on deep learning, and relates to the technical field of image processing, and the method comprises the steps: collecting and preprocessing multi-modal medical image data, carrying out the registration of medical images of different modals, and fusing and optimizing the registered multi-modal images. Carrying out image block segmentation on the optimized fused image, distributing a significance score of each block, marking a candidate lesion region based on the significance scores, and generating a region mask; and carrying out time sequence change analysis on the region mask to generate a multi-time-point focus region heat map, and carrying out visual display. According to the method, deep learning is combined with a traditional registration and fusion technology, the registration precision can be remarkably improved, the fusion effect can be optimized, multi-time-point focus change visual analysis is achieved through a dynamic focus heat map generation technology, and accurate and dynamic support is provided for diagnosis and treatment of complex diseases.

The invention discloses an intelligent detection method and device for fusing a medical image learning image, and relates to the technical field of medical image processing. The method comprises the following steps: acquiring and preprocessing a bimodal medical image, and extracting a feature map through multi-scale decomposition; constructing a cross-modal correlation model, and setting a modal attention mechanism (embedding anatomical structure prior guidance feature complementation) and a morphological attention mechanism (setting lesion morphological constraint weight); the method comprises the following steps: collecting multiple types of image samples, pairing according to a focus form and an imaging mode to construct a bimodal joint data set, and correlating and labeling to generate a training data set with modal attributes; after a multi-stage iteration training model, inputting the preprocessed image to carry out feature fusion so as to obtain a fused image; and generating a lesion probability graph according to the fused image, positioning a lesion area through multi-threshold segmentation, and outputting a detection result. The system comprises a data acquisition module, a preprocessing module and the like. The method improves the accuracy and reliability of medical image detection, and is suitable for clinical multi-modal image analysis.

The invention discloses a plateau mountain road disaster identification method and system based on multi-source remote sensing image restoration and super-resolution reconstruction. The method comprises the following steps: acquiring and preprocessing a multi-source remote sensing image of a plateau mountain region, and extracting landform measurement parameters based on a digital elevation model; a super-resolution reconstruction network fusing deformable convolution and Transform is constructed, and a low-resolution image is reconstructed by using constraint training of a composite loss function containing geomorphic measurement parameters; performing feature extraction and adaptive weighted fusion on the preprocessed image and the reconstructed high-resolution image; based on the fused image, utilizing a multi-task deep learning model to identify landslide, debris flow and roadbed subsidence disasters along the highway; and carrying out morphological optimization and boundary refinement under GIS constraint on an identification result, and outputting a disaster thematic map. According to the invention, the precision and reliability of road disaster identification in a complex terrain environment are effectively improved.

The invention discloses a fusion processing method based on multi-modal oral image data, and relates to the technical field of image data processing, and the method comprises the steps: carrying out the rigid registration of a structural mask, obtaining an alignment parameter, and carrying out the non-rigid registration of a gray image, and obtaining an enhanced gray image; inputting the structure mask, the alignment parameter and the enhanced gray level image into a CycleGAN model to obtain a multi-modal registration fusion image; performing wavelet transform on the multi-modal registration fusion image to obtain multi-scale frequency domain data; taking data acquired at different times as time sequence image data; performing space-time alignment to obtain a multi-modal space-time registration data set; performing feature decoupling based on the multi-modal space-time registration data set to obtain layered features; and based on the hierarchical features, the structure mask and the gray level image, carrying out regional adaptive fusion to obtain a multi-modal optimization fusion image. The technical effect of improving the multi-modal fusion precision and efficiency is achieved.

The invention relates to the technical field of medical image processing, in particular to an engineered heart tissue pulsation trajectory automatic analysis method based on an image recognition algorithm, which comprises the following steps: S1, multi-modal image fusion: performing space-time registration and feature fusion on acquired multi-modal heart images to generate a fused image sequence; s2, cardiac muscle tissue segmentation: outputting a cardiac muscle tissue segmentation result with a timestamp; s3, motion track modeling: generating three-dimensional track point cloud data in a pulsation period; s4, feature parameter extraction: performing spatial-temporal feature analysis on the track point cloud data, and extracting multi-dimensional motion parameters; and S5, heterogeneity atlas generation: generating a cardiac pulse heterogeneity atlas according to the multi-dimensional motion parameters. According to the method, automatic analysis of the cardiac pulse track and generation of the heterogeneity atlas based on the multi-modal image and space-time modeling are realized, and the precision and the intelligent level of cardiac motion anomaly recognition are remarkably improved.

The invention discloses a brain tumor multi-mode large model construction method, device and equipment and a storage medium, and is applied to the technical field of brain tumor imagines.The method comprises the steps that pixel-concept level alignment is conducted on a multi-mode MRI image and a pathological text; constructing a multi-modal feature fusion network for fusing image features and text features by adopting an attention mechanism of pathology perception and combining medical semantic information; training the multi-modal feature fusion network to generate an analysis report and a segmentation result; according to the technical scheme of multi-task cooperation, cross-modal pathological semantic accurate alignment, pathological knowledge graph injection and lightweight and continuous optimization parallelization, full-process coverage of brain tumor accurate segmentation, analysis report generation and prognosis prediction is achieved, the problems that a traditional model lacks pathological semantic support and is insufficient in clinical adaptability are solved, and the clinical adaptability of the traditional model is improved. And the deployment feasibility and the dynamic optimization capability are also considered.

The invention provides a vascular calcification analysis method and device based on a non-enhanced CT image, and relates to the technical field of artificial intelligence. The method comprises the following steps: analyzing scanning parameter information of a non-contrast enhanced scanning image set; standardizing the non-contrast enhanced scanning image set based on the scanning parameter information to obtain a preprocessed image sequence; inputting the preprocessed image sequence into a multi-modal registration network, and generating a fused image based on the obtained key anatomical mark points; inputting the fused image into an attention-enhanced segmentation network, and determining blood vessel segmentation masks of the thoracic aorta, the common carotid artery and the intracranial artery; a candidate voxel set is extracted from the fused image based on the blood vessel segmentation mask, three-dimensional connected domain analysis is carried out on the candidate voxel set, and a blood vessel calcification three-dimensional area is determined; and inputting the vascular calcification three-dimensional region into the multi-task prediction model, determining a quantitative score of the vascular calcification degree, and improving the accuracy and reliability of a vascular calcification evaluation result.

The invention provides a clinical test result prediction model driven by multi-modal data, which relates to the technical field of digital data construction and comprises a data acquisition and preprocessing module, a feature extraction and fusion module, a self-adaptive prediction model, a feedback and updating module and a visualization module. By fusing multi-modal information such as images, genomes, laboratory data and text records, the scheme can more comprehensively obtain multi-dimensional health data of the patient. The feature dynamic weight distribution module dynamically adjusts the weights of different modal data based on global and local attention mechanisms, and ensures the prominent contribution of key features in different prediction scenes, thereby significantly improving the prediction precision and stability of the model.

The invention relates to a reconstruction method of a liver, gall and pancreas three-dimensional image. According to the technical scheme, the method comprises the following steps: introducing a segmentation network for U-Net variants of the liver, the bile duct and the pancreas, and segmenting a multi-modal image by using a convolutional neural network, including separation of the liver, the bile duct and the pancreas; a boundary enhancement module is introduced in the segmentation process; performing spatial alignment on the image data of different modalities by adopting a deep learning driven non-rigid registration algorithm; carrying out registration by utilizing anatomical feature points of the organs; according to a registration result, performing multi-modal fusion on the liver image, the bile image and the pancreas image, and constructing fusion image data by comprehensively utilizing the bone and tissue contrast ratio of CT, the soft tissue resolution ratio of MRI and the real-time dynamic characteristics of ultrasound; extracting contours of liver, gall and pancreas from the segmented image, and constructing a preliminary three-dimensional mesh model; and according to the geometrical shapes of different organs, adaptive adjustment is carried out, and the distribution of grids during reconstruction is controlled to be matched with the organ shapes.

The invention discloses a liver and gall lesion early screening system and method based on image fusion, and relates to the technical field of medical image processing and computer-aided diagnosis, and the method comprises the following steps: reconstructing a multi-modal image space-time coordinate system under a unified event time baseline, generating a respiratory displacement field and a magnetic sensitive pulse fingerprint, and constructing an artifact suspicion map; and performing anti-fact playback based on the artifact suspicion chart, performing frame-by-frame playback on the image acquisition sequence, quantifying artifact superposition tracks with consistent directions, and solidifying an artifact anchor point set. According to the method, space-time coordinates are constructed based on a unified event time baseline, a breathing displacement field and magnetic sensing pulse fingerprints are introduced, anti-fact playback, distortion kernel inference and residual decoupling are combined, artifact recognition and fusion intervention are achieved, and artifact closed-loop elimination is completed by judging threshold-driven fusion regulation and time reversal phase gating, so that the artifact recognition accuracy is improved. And the fused image authenticity is improved.

The invention discloses a cold region saline-alkali soil rice growth assessment method and system based on image data fusion, and relates to the technical field of rice growth monitoring. According to the main technical scheme, the method comprises the steps of obtaining multi-source fusion image data, environment monitoring data and actually measured growth indexes corresponding to different growth stages of rice in a cold region saline-alkali soil region in a historical period; performing deep feature extraction on the multi-source fusion image data to obtain key image features; performing deep feature extraction on the environment monitoring data to obtain key environment features; and respectively processing the key image features and the key environment features corresponding to different growth stages in a historical period by using an attention mechanism, and constructing a growth evaluation model corresponding to the rice in the cold region saline-alkali soil region based on the processed key image features, the processed key environment features and actually measured growth indexes. The main purpose is to improve the accuracy of rice growth evaluation of the cold region saline-alkali soil.

The invention relates to the technical field of medical image analysis, and discloses an HER2 positive breast cancer auxiliary prediction method based on multi-modal data fusion. According to the method, preoperative MRI, clinical pathology data and RNA-seq data meeting quality control standards are obtained through dynamic case screening, traditional image features, multi-scale Hessian matrix blood vessel-structure features and topological persistent coherent curvature features are extracted through multi-plane segmentation, and key features are obtained through three-level screening. A multi-factor logistic model fusing radiomics, clinical features and immune scoring is further constructed, a pCR probability value and a feature contribution heat map are output, and MRI data quality control, real-time analysis and interactive report generation are achieved through a matched computer system. The method breaks through the limitation of single modal analysis, dynamically associates the image features with the molecular mechanism, and provides interpretable decision support for precision medical treatment.

The invention relates to an urban scene classification method based on a high-resolution multi-view remote sensing image, and belongs to the technical field of intelligent extraction of urban scene information. The method comprises the following steps: acquiring satellite remote sensing image data including multispectral and multi-view images, and preprocessing the data; carrying out pixel value extraction on the multi-view image to construct multi-angle three-dimensional features; establishing an LCZ sample set for a local climate region LCZ scene classification system; an MSVTNet city scene classification model fusing the attention and vision Transform of the multi-scale pyramid is established; and training the city scene classification model, and performing LCZ scene classification on the real-time remote sensing image after the training is completed. According to the method, the height difference change and the vertical structure feature of the ground object are extracted by using the multi-view image, the plane feature can be effectively supplemented, the proposed network model can effectively extract and fuse the global-local feature of the image, and the scene classification precision is remarkably improved.

The invention provides a laryngeal cancer multi-mode prognosis prediction method and a laryngeal cancer multi-mode prognosis prediction system fusing a CT (Computed Tomography) image and a ViT model. Relates to the technical field of biomedical images. The method comprises the following steps: acquiring and preprocessing multi-modal data of a laryngocarcinoma patient; carrying out lightweight compression, redundant information screening and robustness training on the ViT model to obtain an optimized ViT model; extracting depth features of the CT image data based on the optimized ViT model, and performing multi-stage fusion on the depth features and clinical and genome data to construct a prognosis prediction model; and performing risk stratification on the patient according to a prognosis prediction result predicted by the prognosis prediction model, and outputting treatment guidance suggestions based on the risk stratification. Through ViT model optimization, multi-modal data fusion and clinical adaptation design, precise prediction and personalized treatment guidance of laryngocarcinoma prognosis are realized, and the problems of insufficient image degradation processing, low model deployment efficiency and the like in existing laryngocarcinoma prognosis prediction are solved.

The invention relates to the field of remote sensing image processing and computer vision, and discloses a satellite image cascade matching method and system based on double-branch context awareness, and the method comprises the following steps: constructing a heterogeneous double-branch feature extraction network, extracting multi-scale detail features through a main feature network, and capturing global semantic information through a context coding network; fusing multi-scale main features and context features to construct a group-related cost body, and adaptively enhancing key region matching response in combination with a channel incentive mechanism; optimizing the cost body by adopting a cross-scale information transfer and multi-dimensional attention fusion strategy, and generating a multi-scale initial disparity map; image edge geometric features and semantic contexts are fused, and high-resolution parallax details are recovered through a multi-stage residual decoder. According to the method, the problem of matching fuzziness of a traditional method in weak texture, repeated structure and parallax abrupt change areas of a satellite image is solved, and the three-dimensional reconstruction precision and the edge detail integrity in a complex scene are remarkably improved.

The invention discloses a breast cancer prognosis real-time evaluation system and method fusing a multi-modal image and a co-disease network, and relates to the technical field of breast cancer prognosis evaluation. According to the system, a comprehensive feature matrix fusing images, genes and clinical features is constructed by acquiring a mammary gland medical image, extracting focus features and combining gene expression information of a focus area and co-disease data of a patient, feature weighting is carried out based on an attention mechanism, and finally a prognosis risk score is output by utilizing a multi-layer perceptron model. And risk grading and intervention suggestion generation are realized. According to the system, a co-disease network and a gene interaction network are introduced for modeling, relevance expression among multi-source information is enhanced, and pathological features and systematic health status of patients can be reflected comprehensively. The intelligent level of prognosis evaluation can be improved, and good clinical popularization value is achieved.

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 discloses a CT image analysis method and system based on a neural network, and relates to the technical field of CT image analys.The method comprises the steps that an original CT image is obtained after user authorization, a Laplace operator is adopted to strengthen a focus boundary, and a circular region of interest is intercepted to remove edge sensitive information; extracting edge and texture information in the standardized image; focus area features are focused step by step; executing characteristic distillation balance based on category sample distribution, and outputting a focus characteristic graph with local perception enhancement and sample balance characteristics; segmenting the lesion feature map into serialized units, embedding position codes, inputting the serialized units into a plurality of layers of encoders, and fusing an image structure and text indication information through a dynamic adjustment mechanism; performing linear classification on the global semantic vector to output a diagnosis result, generating a focus thermodynamic diagram, and superposing the focus thermodynamic diagram to an original image for visualization; and performing dynamic optimization based on doctor feedback. The accuracy of feature analysis is improved; the overall operation efficiency of the system is improved.

The invention relates to the technical field of neurosurgical intelligent operations, and discloses a neuroendoscopic surgery intelligent puncture positioning system based on image navigation, and the system comprises a preoperative modeling module which constructs a preoperative tissue model and a corresponding elastic tensor field based on a three-dimensional medical image of a patient; the image fusion module is used for receiving the data of the preoperative modeling module, collecting intraoperative endoscopic images, predicting the tensor disturbance of tissues through a deep learning model, and feeding back the updated tensor disturbance to the tensor updating module; and the tensor updating module is in two-way communication with the image fusion module and is used for superposing tensor disturbance obtained by processing the intraoperative image with the preoperative tensor. Through fusion of an image navigation technology and preoperative modeling, a three-dimensional positioning frame based on neuroendoscopy operation requirements is constructed, intelligent planning and real-time correction of a puncture path are realized, important brain regions and vascular structures are effectively avoided, intra-operative risks are reduced, and positioning precision and safety are improved.

The invention relates to the technical field of medical images, in particular to a neural CT map segmentation method based on high-precision registration. The method comprises the following steps: acquiring a nerve CT image; performing coarse registration based on mutual information according to the CT image, and performing multi-scale fusion based on wavelet to obtain a fused image; performing standard map preprocessing on a preset standard nerve CT map database according to the fused image to obtain a preprocessed standard map set; adding a multi-scale feature fusion module to a pre-established basic U-Net architecture, and embedding a space attention module to obtain a to-be-trained U-Net network; and performing deep registration network training on the to-be-trained U-Net network by using the fused image and the preprocessed standard atlas set to obtain a trained U-Net network. According to the invention, through technologies of multi-modal image fusion, deep learning registration and conditional random field optimization, a segmentation effect with higher precision and higher robustness is realized.

The invention proposes a thoracic surgery intelligent planning system based on multi-modal image fusion, and relates to the technical field of surgery planning, and the system comprises an image processing module which is used for receiving multi-modal medical image data, carrying out the preprocessing and cross-modal registration of the medical image data, and generating fused image data; the three-dimensional reconstruction module is used for segmenting the key anatomical structure in the thoracic cavity based on the fused image data and constructing a three-dimensional model; the operation planning module is used for generating a candidate operation path scheme according to a multi-objective optimization algorithm based on the three-dimensional model; the interaction adjustment module is used for determining a final operation path scheme and updating a risk assessment result in real time; and the report generation module is used for generating a surgical planning report. The method can solve the problems that the existing multi-modal image registration precision is insufficient, the sensitivity of three-dimensional reconstruction to a tiny anatomical structure is low, and the surgical path planning is lack of dynamic risk assessment and real-time interaction adjustment capability, and realizes the generation and optimization of a high-precision and high-safety thoracic surgery planning scheme.

The invention specifically discloses an automatic generation method for a cervical vertebra disease rehabilitation prescription, and the method comprises the steps: constructing a cervical vertebra image feature automatic measurement system based on computer vision and deep learning, so as to monitor the training motion quality of a patient in real time, and obtaining the kinematics parameters of rehabilitation exercise training; acquiring a medical image and clinical data of a patient, and standardizing the medical image to construct a standard space of a multi-modal medical image; performing automatic segmentation and anatomical structure calibration on a vertebral body-intervertebral disc based on the standard space to obtain fusion image features; and constructing a multi-modal hierarchical decision model fusing the image features, the kinematics parameters and the clinical data to carry out quantitative analysis on pathological feature parameters, and outputting a standardized illness state assessment result conforming to international clinical guidelines. The dynamic optimization of the treatment scheme can be realized through a feedback mechanism, and the treatment effect is improved.

The invention relates to the field of remote sensing image processing and geological disaster monitoring, in particular to a goaf three-dimensional subsidence basin reconstruction system based on multi-source remote sensing image fusion, which comprises a heterogeneous data preprocessing module, a differential geometry-based data fusion module, a three-dimensional subsidence basin reconstruction module, a subsidence dynamic monitoring module and a virtual reality interaction module, the system innovatively introduces a differential geometry theory to solve the problem of heterogeneous fusion of a high-resolution optical satellite image, an SAR satellite image and airborne remote sensing Lidar point cloud data, maps multi-source data to a unified feature space through manifold learning, calculates an adaptive fusion weight through curvature analysis, constructs a multi-scale feature correlation matrix based on a geodesic line distance, and achieves the fusion of the high-resolution optical satellite image, the SAR satellite image and the airborne remote sensing Lidar point cloud data. A high-precision fusion image is generated, three-dimensional reconstruction is carried out in combination with subsidence basin geological parameters and a physical constraint method, it is ensured that a reconstruction result conforms to an actual subsidence physical rule, dynamic monitoring and virtual reality interaction of the subsidence process are achieved, and parameter adjustment and model optimization are supported.

The embodiment of the invention discloses a cervical cancer wettability risk assessment method and system based on three-dimensional ultrasound and tomographic ultrasound imaging. The method comprises the steps that three-dimensional ultrasound image data and tomographic ultrasound imaging data of a target cervical region are acquired; carrying out preprocessing and image registration on the three-dimensional ultrasonic image data and the tomography ultrasonic imaging data; fusing the spatial structure information and the chromatography tissue information by adopting a weighted fusion algorithm to generate fused image data containing multi-dimensional features; performing feature extraction on the fused image data by using a feature extraction model, and combining morphological features and histological features to generate a high-dimensional feature vector; extracting hemodynamic parameters based on the three-dimensional ultrasonic image data; inputting the high-dimensional feature vector and the hemodynamic parameters into a risk assessment model to generate a cervical cancer wettability risk score; according to the risk score, outputting a cervical cancer wettability risk assessment result; and high-precision cervical cancer wettability evaluation is realized.

The invention discloses a surface feature information extraction method based on a hyperspectral and high-resolution remote sensing image, and relates to the technical field of remote sensing. The method comprises the following steps: acquiring a hyperspectral and high-resolution remote sensing image, and preprocessing the hyperspectral and high-resolution remote sensing image; cutting and fusing the preprocessed image to obtain a fused image; performing spectral analysis and wave band screening on the fused image to obtain first image data; performing image enhancement processing on the first image data to obtain a plurality of grayscale images; performing wave band synthesis on the plurality of grayscale images, and performing grid analysis and wave band screening to obtain second image data; performing threshold segmentation on the second image data, and performing grid surface turning to obtain a first vector; obtaining DEM data in a preset elevation range of the research area, and performing grid surface turning to obtain a second vector; and cutting the first vector according to the second vector to obtain a distribution vector result. According to the method provided by the embodiment of the invention, the extraction precision and reliability can be effectively improved.

The invention discloses a desert remote sensing image lake extraction method based on SSI and water body index fusion, and the method comprises the steps: calculating a suspended sediment index (SSI) and a water body index based on remote sensing image data obtained by a GEE platform; carrying out fusion processing on the suspended sediment index and the water body index to obtain a fusion image so as to enhance the distinction degree of the lake area and the background area; a classification method is provided through single-index threshold segmentation; and classifying the fused image based on the proposed classification method, and accurately extracting the lake area. The lake extraction accuracy in the desert environment can be effectively improved, especially in a lake area containing high-concentration silt and a desert area with high earth surface reflection. Through the fusion of the two technologies, the advantages of different spectral indexes can be fully utilized, and more accurate water body identification can be realized in a complex environment.

The invention provides a pneumoconiosis image multi-modal identification method and identification system, and relates to the technical field of image identification. The identification method provided by the invention comprises the following steps: carrying out space mapping, image registration and modal fusion on a lung CT image and a corresponding lung DR image to obtain a fused image; performing image amplification on the fused image in the generative adversarial network to obtain a data set; training the initial model based on the data set to obtain a recognition model; and recognizing the to-be-recognized fused image based on the recognition model to obtain a recognition result. According to the invention, the identification accuracy of the pneumoconiosis image can be improved.

The invention belongs to the technical field of remote sensing image processing, and discloses a remote sensing image space-time fusion method and system, and the method introduces a Mama module in a space domain and a frequency domain, so as to enhance the extraction and modeling capability of multi-scale features and structural information. In a spatial domain, the Mama module is used for capturing a local structure and geometric details of an image; in a frequency domain, for a high-frequency sub-band after wavelet transform, a Mama module is introduced to carry out feature extraction, so that the context dependency relationship and structural features of high-frequency information are more fully mined. Double-domain interactive fusion is realized through a channel attention and space attention mechanism, the space definition and spectrum consistency of the fused image are improved, the method has strong generalization and adaptability, space domain and frequency domain information is fully utilized, and the precision and efficiency of obtaining the remote sensing image are improved.

The invention discloses a three-dimensional brain network dynamic segmentation method based on deep learning. The method comprises the following steps: S1, constructing fused image volume data with consistent time and consistent space; s2, a multi-scale sparse Transform coding feature pyramid is generated; s3, obtaining a same-scale brain region map structure; s4, taking the cross-scale node alignment fusion graph structure as initial output of a cross-scale node alignment fusion mechanism; s5, obtaining a first round of fusion brain region graph node embedding feature; s6, an updated multi-scale sparse Transform coding feature pyramid is obtained, and the step S3 to the step S5 are repeated until interaction updating of the multi-scale sparse Transform-GNN is completed; and S7, obtaining a three-dimensional brain region dynamic segmentation result. According to the method, collaborative extraction of local fine granularity and global coarse granularity features is realized, redundant information can be effectively inhibited, and the modeling capability of spatial structure and functional connection features in a cross-modal fusion image can be enhanced.

The invention discloses a deep learning river ice extraction method based on an optical and SAR fusion image, and belongs to the technical field of remote sensing geoscience application. The method comprises the following steps: acquiring an optical image and an SAR image which are in the same area and contain river ice distribution, preprocessing the optical image and the SAR image, marking river ice areas of the preprocessed optical image and the preprocessed SAR image, and constructing a training data set; a deep learning extraction model is constructed, and the training data set is used for training; and obtaining an optical image and an SAR image of a to-be-extracted river ice distribution region, obtaining a binary image by using the trained deep learning extraction model, converting the binary image into vector data, removing a misjudged river ice region, obtaining a river ice region in the region, and completing river ice extraction. According to the method, advantages of optics and SAR images are combined, an attention mechanism and a multi-level feature fusion strategy are constructed, and technical support is provided for refined river ice drawing in the cold and cold mountainous area.