338 results about "Spatial domain" patented technology

This application relates to the field of artificial intelligence technology and discloses a semantic segmentation system and method for satellite remote sensing images. The system includes: an encoding module configured to extract multi-scale features of the target image; a feature fusion module configured to fuse the original feature maps at various scales through linear attention calculation to obtain a first fused feature, and then split it to generate multiple mixed feature maps with the same scale as the original feature maps; a collaborative decoding module configured to extract spatial domain features and frequency domain features from the original feature maps and the mixed feature maps of the same scale, and perform cross-attention calculation to iteratively fuse spatial domain information and frequency domain information to generate a first tensor; and an output module configured to process the first tensor to obtain the segmentation result of the target image. This system can improve the recognition accuracy of detailed textures and small targets in remote sensing images, and improve the accuracy of semantic segmentation of remote sensing images.

A terminal is disclosed that includes a receiver that receives beamforming information relating to one or more beamformed Channel State Information—Reference Signals (CSI-RSs), the CSI-RSs corresponding to one or more Spatial Domains (SDs) or Frequency Domains (FDs). The terminal also includes a processor that determines a type of the one or more CSI-RSs, and when the type of the one or more CSI-RS is a Type II, performs a Type II port selection codebook with a beamformed CSI-RS of the one or more CSI-RSs in the SD and FD domains.

Apparatuses, methods, and systems are disclosed for configuring UL-TCI states. One method includes receiving a RRC message configuring a set of UL-TCI states. Each UL-TCI state defines at least one uplink spatial relationship. Each uplink spatial relationship corresponds to a spatial domain transmission filter corresponding to a CSI-RS, a SSB, or an SSB. The method includes receiving a MAC CE message comprising information that activates a subset of the set of UL-TCI states for use in an UL-TCI field of DCI used to schedule a codebook-based PUSCH transmission. The method includes receiving the DCI with the UL-TCI field indicating the at least one uplink spatial relationship for the UE to use for the codebook-based PUSCH transmission. The method includes transmitting the codebook-based PUSCH transmission with the at least one uplink spatial relationship indicated by the UL-TCI field in the DCI.

This invention belongs to the field of image processing and computer vision technology, and discloses a lightweight single-image super-resolution reconstruction method based on a directional stripe hybrid network. It captures directional structural information of the image from both the spatial and frequency domains through horizontal stripes, vertical stripes, local window attention of the directional stripe self-attention module, and direction-specific convolution of the directional high-frequency enhancement module. The directional stripe self-attention module effectively captures long-distance dependencies extending along specific directions; the directional high-frequency enhancement module extracts high-frequency components through frequency domain decomposition, combines them with direction-specific convolution, and introduces a dual-guided adaptive enhancement mechanism. It identifies texture-rich regions through local variance maps and structural boundaries through edge maps, generating adaptive weights to guide feature enhancement, accurately enhancing high-frequency edges and textures with different orientations while effectively suppressing noise. The two work together to achieve spatial-frequency domain dual-level directional modeling, significantly improving the reconstruction quality of images rich in directional textures, such as architectural scenes and cartoon images, while maintaining a lightweight design.

This invention relates to the field of computer graphics data processing technology, specifically a method for hot-plugging dynamic outer contour meshes driven by three-dimensional spatial domain and time slices. The steps include: S1, receiving air traffic control message information in real time via a transmission protocol, converting geographic coordinates into target world coordinates, and simultaneously generating management files; S2, constructing a spatial domain mesh in real time based on the four types of management files using a combination of outer contour extrusion and ear-cutting methods; S3, mapping the spatial domain ID and time slice from the message to the spatial domain mesh constructed in S2, constructing a double-key hash index of spatial domain ID and time slice, which serves as the sole basis for the hot-plug manager to determine whether to update or create a mesh, dynamically executing spatial domain mesh hot-plugging. Employing the contour extrusion-ear-cutting method and vertex cache reuse mechanism, combined with the "spatial domain ID and time slice" double-key hash index, it supports dynamic construction and hot-plugging operations of the mesh model, meeting the low-latency rendering requirements of real-time air traffic control data.

The application discloses a railway catenary foreign matter detection method based on frequency-space cooperation purification, and relates to the field of railway engineering technology.The method comprises the following steps: at the front end of the main network, input shallow feature maps are respectively processed by a spatial domain adaptive compensation branch and a frequency domain dynamic filtering branch, and a purified feature map is obtained through fusion; the bottleneck layer of a feature extraction module in the main network is reconstructed, and linear topological features are extracted by using a multi-scale parallel strip convolution path; a spatial attention map is generated by a shape perception substream for the feature map output by the neck network, pixel-level weighting is performed on the regression substream features, and a bounding box prediction result is output, and the network is trained by using a total loss function comprising a classification loss, a bounding box overlap loss and a shape penalty loss.The application can inhibit regular background noise, improve the continuity of foreign matter feature expression, and help improve the positioning accuracy of small foreign matters.

This application discloses a video frame error repair method, apparatus, device, and medium based on spatial domain filtering, relating to the field of video encoding and decoding technology. The method includes: acquiring a target video frame and determining a reconstructed image block; determining target boundary pixels based on the distance between each pixel in the reconstructed image block and the image block boundary, and adjusting the pixel values ​​of the target boundary pixels to obtain a boundary-adjusted image block; performing spatial domain filtering on the boundary-adjusted image block to obtain a filtered image block; acquiring target texture information from the neighborhood of the filtered image block, and superimposing the target texture information onto the corresponding target boundary pixels in the filtered image block to complete the video frame error repair. By making the filtering intensity in the boundary region lower than that in the non-boundary region, edge details are preserved; and by acquiring texture information from the neighborhood and superimposing it onto the boundary pixels, the problem of lacking natural texture is solved.

The present application belongs to the technical field of electric power construction, and particularly relates to a power transmission and distribution construction monitoring method and system based on air-ground communication. The present application obtains an initial image sequence of a construction site through image acquisition equipment deployed in cooperation with the air and the ground, first performs frequency domain and spatial domain joint adaptive denoising on the image, and then identifies the construction process node in real time; the feature anchor points of the intra-frame static ground features are extracted, rigid alignment is performed through feature anchor point matching, and a standardized image sequence that is spatially aligned throughout the construction period is generated; a monitoring target area that matches the process node is demarcated, the gray scale gradient and edge texture feature vectors in the area are extracted, and are transmitted to the management and control end through an air-ground communication link, and are compared with the standardized construction feature benchmark corresponding to the process, to complete accurate judgment of safe operation. The present application solves the pain points of construction scene image noise and spatial misplacement, and greatly improves the monitoring recognition accuracy and management and control reliability.

This invention provides a method and system for three-dimensional high-frequency controllable source electromagnetic numerical simulation with dual-source excitation. The method constructs a single-source electromagnetic field grid model and a three-dimensional spatial domain total field equation, converting it into a quadratic field control equation by separating the primary and secondary fields. Fourier transform is used to transform the equations to the wavenumber domain, and shape function interpolation is performed iteratively to obtain the wavenumber domain numerical solution. A single-source forward excitation solution is obtained through inverse Fourier transform. A single-source reverse excitation solution is obtained by reversing the current direction. Finally, the forward and reverse solutions are vector-superimposed to obtain the dual-source excitation simulation result. This invention solves the technical problems of existing technologies, such as simulation distortion caused by neglecting displacement current in the 100kHz-10MHz frequency band, numerical calculation interference from field source singularities, and low spatial resolution of complex underground anomalies under single-source excitation. By simulating gradient field response characteristics through dual-source reverse superposition, the sensitivity and exploration resolution of underground target boundaries are significantly improved.

This invention discloses a small target detection method based on frequency-enhanced deformable convolution, comprising the following steps: S1, dividing several small target images from remote sensing scenes according to a preset ratio to obtain a training set and a test set of several remote sensing small target images; S2, inputting the training set into a constructed small target detection network for training, and saving the optimal small target detection network; the small target detection network adopts an adaptive fusion mechanism to fuse spatial domain features and frequency domain features to improve feature representation ability; deformable convolution is used to learn fine target deformation features to improve the accuracy of boundary localization; the fused features are input into a target detection head to complete target classification and bounding box regression; S3, inputting the test set of several remote sensing small target images into the optimal small target detection network to obtain the results of the test set images. This invention can enhance the target detection model's ability to perceive high-frequency details and complex geometric structures.

The application discloses a space network networking method and device, a network resource arrangement platform and a storage medium. The method comprises the following steps: in the case that a UAV service accesses a network, the position of the UAV performing the UAV service is determined; the service score corresponding to the cell within the preset distance range from the position of the UAV is determined; the preset number of cells with the highest service score are determined as the space candidate working cells, and the number of adjustable beams of the space candidate working cells is determined; and the space-to-air beam and the ground-to-air beam of the space candidate working cells are determined and set according to the number of adjustable beams. The application solves the technical problems that the UAV space coverage scheme in the related art mostly needs to add equipment for transmission transformation, the cost is too high, the scheme of covering the space by directly using the existing site resources has too great influence on the ground coverage, and the beam resources for covering the space are wasted.

This invention proposes a wavenumber spectrum reconstruction and rapid fusion imaging method and system for arbitrary bistatic SAR. It proposes a method to consistently center the wavenumber spectrum of the spatial domain signal of the imaging scene, introduces a two-dimensional wavenumber spectrum position correction filter to compress the wavenumber spectrum width of the imaging scene, reduces the Nyquist sampling requirement in the spatial domain, and achieves accurate SAR imaging of the imaging scene with less computation, thus improving the imaging efficiency of the imaging method. Furthermore, while achieving accurate and consistent centering of the wavenumber spectrum of the imaging scene, this invention proposes a new method for calculating the minimum Nyquist sampling requirement of the spatial domain signal of the imaging scene. Based on the framework of the time-domain FFBP imaging method, this invention enables the imaging method to have good parallelism, utilizing parallel processing mechanisms on hardware platforms such as GPUs and FPGAs, which will further significantly improve the imaging efficiency of the time-domain FFBP imaging method in bistatic SAR systems in practical applications.

Provided in the present application are a communication method and a communication apparatus. The method comprises: receiving reference signals on Ks reference signal resources, wherein Ks is an integer greater than or equal to 1; and sending a PMI, wherein a precoding matrix corresponding to the PMI is determined on the basis of spatial domain bases corresponding to K' reference signal resources from among the Ks reference signal resources, wherein K' is an integer greater than or equal to 1 and less than or equal to Ks. By means of the method provided in the present application, communication overheads can be effectively reduced.

The application discloses a plant pest target detection method and system, comprising acquiring a plant image and extracting a multi-scale stem feature; for each target scale, selecting features of adjacent scales, generating cross-layer attribution weights channel by channel and performing selective aggregation to obtain first fusion features; inputting the first fusion features under multiple target scales into a topology-aware spatial refinement module, generating a position-related spatial reorganization kernel according to uplink and downlink connection relationships of a pyramid, introducing a boundary and texture-sensitive detail bypass at the same time, fusing high-frequency information of the reorganization kernel and the detail bypass in a spatial dimension to obtain second fusion features; sending the second fusion features into a detection head for classification regression to output predicted target positions and categories; through coupling of channel domain layered attribution and spatial domain topology refinement, the application realizes simultaneous improvement of high threshold positioning quality, scale robustness and task collaboration.

This invention belongs to the field of remote sensing image processing technology and discloses a method and system for spatiotemporal fusion of remote sensing images. This method introduces Mamba modules in both the spatial and frequency domains to enhance the extraction and modeling capabilities of multi-scale features and structural information. In the spatial domain, the Mamba module is used to capture the local structure and geometric details of the image; in the frequency domain, the Mamba module is introduced for feature extraction of the high-frequency subbands after wavelet transform, thereby more fully exploring the contextual dependencies and structural features of high-frequency information. Through channel attention and spatial attention mechanisms, dual-domain interactive fusion is achieved, improving the spatial clarity and spectral consistency of the fused image. It possesses strong generalization and adaptability, fully utilizing information from both the spatial and frequency domains, and improving the accuracy and efficiency of acquiring remote sensing images.

An electronic device includes one or more processors comprising processing circuitry, and memory comprising one or more storage media storing instructions that, when executed individually or collectively by the one or more processors, cause the electronic device to determine a first spectrum comprising a plurality of frequency components for an image containing noise determined at each of a plurality of levels by transforming the image into a frequency domain at the plurality of levels, determine a plurality of tokens for a plurality of patches by segmenting the first spectrum into the plurality of patches and encoding the plurality of patches based on positions of corresponding patches, denoise the plurality of tokens based on the image, determine a second spectrum obtained by denoising the first spectrum, based on the plurality of denoised tokens, and generate a high-resolution (HR) image with an increased resolution of the image by inversely transforming the second spectrum into a spatial domain at the plurality of levels.

The application belongs to the technical field of remote sensing image processing and deep learning, and relates to a space-frequency domain neural network remote sensing image semantic segmentation method based on Mamba. The method introduces a space domain and frequency domain dual domain attention module, respectively uses a linear attention mechanism for space domain feature representation and an attention mechanism for frequency domain representation; a boundary perception head is designed to enhance the boundary representation with graphic information features in the feature extraction stage, and the perception ability of the model to the boundary is improved; and a supervised contrast learning strategy is combined to enhance the aggregation ability of semantic information in the feature representation in the decoder stage. Through the synergistic effect of the above modules, the application effectively overcomes the problems of single optimization method and fuzzy boundary segmentation of the existing Mamba model without modifying the hardware optimization code and increasing the model parameter quantity, and significantly improves the segmentation accuracy and running efficiency of the remote sensing image semantic segmentation model.

The application discloses a satellite large-scale MIMO beam structure downlink precoding method and uplink detection method. In the beam base channel model established by the application, the scaling direction cosine is uniformly sampled and quantized, and the uplink / downlink beam matrix is constructed, the space domain channel vector is expressed as the product of the equivalent uplink / downlink channel gain, the uplink / downlink beam matrix and the diffusion vector. Based on the downlink / uplink beam base channel model, the statistical channel information is used to design the downlink precoder / uplink detector under the beam structure, and the high-dimensional space domain precoder / detector design is converted into the low-dimensional beam domain precoder / detector design. Using the designed precoder and detector, low-complexity downlink precoding transmission processing and uplink receiving processing are respectively carried out. The application fully utilizes the sparse characteristics of the satellite channel, guarantees the uplink and downlink ergodic and rate performance of the satellite large-scale MIMO system, and effectively reduces the design and implementation complexity of the precoder and the detector.

This application discloses an image feature extraction method, device, and storage medium. The image feature extraction method includes: extracting frequency domain information of an image to be processed; dividing the image to be processed into regions in the spatial domain based on the frequency domain information to obtain low-frequency image regions and other frequency domain image regions; removing interfering image information from the other frequency domain image regions to obtain frequency domain image templates corresponding to the other frequency domain image regions; performing attention calculation on the other frequency domain image regions and the low-frequency image regions based on the frequency domain image templates to obtain attention features corresponding to the other frequency domain image regions; and fusing the attention features and the image features corresponding to the low-frequency image regions to obtain the image feature extraction result. A morphological attention mechanism can be embedded into a neural network model. When the frequency domain features change, the morphological attention will also adapt accordingly, enabling joint optimization of model parameters and morphological attention processing to accurately extract image features.

The application relates to the technical field of image processing, and discloses a light-weight single-pixel imaging method based on frequency domain-space domain modulation attention, which comprises a single-pixel image data generator and a calculation reconstruction generator; an initial inversion estimation is obtained through inverse mapping of a measurement matrix; then, the initial inversion estimation is input into a dual-domain collaborative attention deep reconstruction network; the network is composed of a plurality of dual-domain collaborative attention module cascades connected in series, a multi-head spectrum calibration module in the module realizes long-distance information reconstruction through dual-domain feature information fusion interaction; a multi-scale-depth dynamic convolution fusion module extracts features in a space domain, accurately processes high-frequency edges, and significantly enhances the reconstruction capability of texture details; and the features are corrected in combination with a residual error guiding update module. Through the dual-domain collaborative mechanism, the reconstruction precision and the detail recovery capability of single-pixel imaging are effectively improved while the network parameter quantity and the calculation complexity are greatly reduced, and the application is suitable for edge deployment.

This invention discloses a method for determining exponential distribution delay based on light intensity and event threshold control, comprising the following steps: Based on the fact that DVS image sensor events exhibit exponential distribution characteristics in the time domain when image information changes rapidly in a scene, an exponential distribution model is established that correlates the time constant with light intensity and a threshold; according to the exponential distribution model and the parameters calculated by the model, the exponential distribution delay of the DVS image sensor event to be analyzed in the time domain is determined, which is characterized by the event trigger rate, including the delay magnitude and its distribution in the spatial domain. This invention, based on the exponential distribution delay model, can quickly and conveniently process and calculate the exponential distribution delay of pixels, providing a basic explanation for the origin of motion artifacts in low-light conditions, and assisting researchers in developing artifact suppression methods based on delay mechanisms.

The application discloses a kind of indoor mobile robot dense mapping and autonomous navigation method based on depth camera, belong to robot simultaneous localization and mapping, robot navigation field.The application uses depth camera and based on interframe matching to image ORB feature to realize real-time positioning to robot, by fusing color image and depth image, and in order to eliminate redundant video frame, key frame extraction method in space domain is introduced, real-time dense three-dimensional point cloud map construction is realized, and it is converted into octree map format and grid map format suitable for navigation, then by ROS Navigation function package, it is combined with existing mobile robot navigation method, realizes indoor mobile robot autonomous navigation scheme based on pure vision scheme.

This invention discloses a multi-scale, cross-domain interactive network for image tampering localization. Addressing the problems of existing methods that suppress semantic information to highlight forensic features, leading to the loss of key context and insufficient generalization ability, this invention constructs a multi-scale, cross-spatial, and noise-domain bidirectional interaction mechanism between semantic features and forensic features. Specifically, the network introduces a bidirectional cross-attention and adaptive gating fusion module, enabling high-level semantic information to guide the discovery of low-level tampering artifacts. Simultaneously, low-level forensic inconsistencies correct high-level semantic understanding, forming a closed-loop learning paradigm where semantics and forensic clues mutually reinforce each other. This method does not rely on specific target semantics and can effectively capture general patterns of image consistency violations, thus exhibiting excellent localization accuracy and robust generalization performance in both traditional editing and AI-generated tampering scenarios.

The present application discloses a method for reconstructing the non-uniform stress field of complex components based on boundary segmentation and frequency-domain bridging, which relates to the technical field of stress field reconstruction. The method includes: first, obtaining the geometric model of the target complex component, identifying the degree of non-uniformity of the stress distribution corresponding to the geometric boundary features and dividing it into multiple sub-regions; then, aiming at the spatial frequency characteristics of the stress distribution in each sub-region, constructing a physics-informed neural network embedded with the elastic mechanics mechanism equation, and training to obtain the stress field sub-network model of each sub-region; then, transforming the spatial-domain stress field output by the sub-network to the frequency domain, establishing a frequency-domain dynamic link between sub-regions through a frequency-domain bridging module, and reconstructing the overall non-uniform stress field of the component through inverse transformation; finally, through a variable fidelity cascaded neural operator network, gradually fusing multi-source stress field data to complete the improvement of fidelity. The method of the present application takes into account both the solution efficiency and the multi-scale stress fitting accuracy, and eliminates the boundary discontinuity problem of segmentation and splicing.

The present application belongs to the technical field of image recognition, and particularly relates to a liquid crystal screen defect detection method based on image analysis. The method comprises: acquiring a liquid crystal screen surface image frequency domain signal and wavelet packet decomposition, and extracting a low frequency and medium-high frequency coefficient matrix; performing convolution on the low frequency coefficient matrix with an adaptive morphological filter kernel, fitting to generate a uniform illumination reference matrix; after weighted correction and denoising of the medium-high frequency coefficient matrix, inverse wavelet packet transformation is performed on the medium-high frequency coefficient matrix and the uniform illumination reference matrix to generate a reconstructed difference image; the local pixel gradient direction vector and the gray level co-occurrence matrix contrast feature of the reconstructed difference image are input into an isolation forest classifier to output a defect position and category. The present scheme shifts illumination separation to the frequency domain, avoids damage to defect pixels by spatial domain filtering, eliminates non-uniform illumination interference, retains the edge gradient and texture structure of small defects, and solves the problem of missing detection of small dark spots under a gradual illumination background.

The application discloses a kind of self-adapting low-light image enhancement method, system, terminal equipment and medium, it is related to image processing technical field.The method includes: obtaining to be enhanced image, determines each pixel brightness value;Brightness value is handled by parameterization processing function, generates complementary information;Complementary information and brightness value are used to construct intuitionistic fuzzy image to express the uncertainty of pixel value;Based on the global statistical characteristics of image, contrast enhancement is carried out to intuitionistic fuzzy image, and preliminary enhanced image is obtained;With the optimization target of maximizing the information content or perceived quality of enhanced image, the adjustable parameter of parameterization processing function is iteratively optimized, and the optimal parameter enhanced image is obtained;It is converted from fuzzy domain back to spatial domain, and the final enhanced image is obtained.The application can improve the contrast of low-light image while retaining structural details, enhance visual quality, and has strong adaptability and high computational efficiency, and can be widely applied to imaging system under complex lighting conditions.

Apparatuses, systems, and techniques to generate blue noise masks for real-time image rendering and enhancement. In at least one embodiment, a vector-valued noise mask is generated and applied to one or more images to generate one or more enhanced images for image processing (e.g., real-time image rendering). In at least one embodiment, the noise mask includes vector values per pixel and is able to handle the temporal domain (e.g., add time to the spatial domain) to improve image quality when rendering images over multiple frames.