21 results about "Compressive imaging" patented technology

The present invention provides a self-adaptive and saliency-aware snapshot compressive imaging (SCI) system. In comparison with the existing SCI systems, the self-adaptive and saliency-aware SCI system of the present invention integrates saliency detection, which feedbacks to the coding masks with a calculated sampling efficiency for updating the coding masks. As such, self-adaptation of the system is achieved, and high-level information such as saliency is also obtained and given consideration, thereby producing reconstruction results with better quality, lower power cost and higher efficiency.

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.

The application relates to the technical field of spectral imaging, and discloses a multi-snapshot hyperspectral compression imaging system. The multi-snapshot hyperspectral compression imaging system disclosed by the application comprises an optical lens, a pixel filter, a surface array camera and a displacement actuator. The displacement actuator moves the pixel filter according to a preset movement mode so that a detector collects multiple image data with the filter being at different positions and having different filtering effects. An upper computer carries out spatial combination through a spectral domain reconstruction neural network model to reconstruct the multiple image data into spatial-spectral data cube reconstruction data. At this time, image processing is carried out through a set original image acquisition device, so that the spatial-spectral data cube reconstruction data reconstructed through the above scheme has high accuracy, thereby solving the technical problem of low image reconstruction accuracy of the existing hyperspectral compression imaging system.

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.

The invention discloses a hyperspectral compression imaging method and system, a terminal and a storage medium in the technical field of hyperspectral image processing, and aims to solve the problem of low reconstruction precision caused by loss of high-frequency details, weak spectral correlation and poor anti-noise capability in a compression imaging process in the prior art. The method comprises the following steps: acquiring compression measurement and a sensing mask of a coded aperture snapshot spectral imaging system, and acquiring an initial feature map according to the compression measurement and the sensing mask; inputting the initial feature map into a pre-constructed U-Net model to obtain a decoded feature map; performing convolutional mapping on the decoded feature map, and obtaining a reconstructed hyperspectral image in combination with the initial feature map; according to the method, long-sequence data are efficiently processed through the spectrum-space state synchronizer, local and global features of the image can be effectively captured through multi-scale decomposition of the wavelet variance modulation block, and the problem that reconstruction precision is not high due to loss of high-frequency details, weak spectrum correlation and poor anti-noise capacity in the traditional compression imaging process is solved.

The invention relates to a high-sensitivity streak camera ultrafast scanning image enhancement and noise reduction method, and belongs to the technical field of ultrafast imaging and signal processing. The method mainly comprises the following steps of: performing spatial modulation on three-dimensional image data coded by a spatial coding element through a periodic grating; carrying out compression imaging on the modulated three-dimensional image data by the streak camera to form a two-dimensional observation image; an ADMM-TV algorithm is adopted to carry out image decoding, and a three-dimensional image containing a periodic modulation structure is recovered; after the three-dimensional image is reconstructed, effective fringe information in the modulated image is further extracted, high-frequency modulation residues are removed, and denoising of the interference fringe image is achieved. Physical separation of signals and noise is achieved through periodic grating modulation, high-frequency interference such as space charge noise can be restrained, the signal-to-noise ratio and contrast ratio of images are improved, and accuracy and stability of diagnosis of an ultrafast imaging system of a VISAR speed interferometer are improved.

This invention discloses an underwater target detection device and operating method based on forward-looking sonar partitioning, relating to the field of underwater acoustic signal processing technology. This invention, by displaying the measurement range, divides the beam domain into grids from two dimensions: beam and range, forming multiple local detection grids. It compresses and images the corresponding beam domain data to form image domain data, and determines the detection threshold for each local detection grid based on the signal-to-noise ratio of the corresponding image domain data to identify underwater targets. By dividing the beam domain data of the detection area into grids according to beam and range, this invention can reasonably divide the beam domain data generated by underwater targets, underwater reverberation, and acoustic ray changes in the beam domain data corresponding to the signal domain. It compresses and images the beam domain data of each local detection grid, and accurately identifies the parameters corresponding to underwater targets, underwater reverberation, and acoustic ray changes from the perspective of obtaining the detection threshold from the signal-to-noise ratio of the image domain data, thus accurately identifying underwater targets.

The invention relates to the technical field of image processing, and discloses a lightweight single-pixel imaging method based on frequency domain-space domain modulation attention, which comprises the following steps: constructing a single-pixel image data generator and calculating a reconstruction generator; initial inversion estimation is obtained through reverse mapping of the measurement matrix; sending the data to a double-domain collaborative attention depth reconstruction network; the network is formed by cascading a plurality of stages of double-domain collaborative attention modules which are connected in series, and multi-head spectrum calibration modules in the modules realize remote information reconstruction through double-domain feature information fusion interaction; the multi-scale-depth dynamic convolution fusion module performs feature extraction in a spatial domain, accurately processes a high-frequency edge and remarkably enhances the reconstruction capability of texture details; and a residual error is combined to guide the updating module to correct the features. Through a double-domain cooperation 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 method is suitable for edge deployment.

This invention specifically relates to a single-exposure compressed imaging reconstruction method and system that integrates pre-trained language models, solving the problem that existing technologies can only decode a single imaging modality. It employs a model framework of "physical constraints-semantic guidance-sparse controlled diffusion," uniformly mapping cross-modal compressed data to align the data stream. On one hand, it utilizes a lightweight mask-gated constraint feature extraction module to extract high-fidelity underlying physical features; on the other hand, it cascades a pre-trained multimodal large language model to generate a global semantic representation. In the controlled diffusion large model denoising and reconstruction stage, an asymmetric sparse controlled DiT architecture is used to solve the problem of easily losing high-frequency details, and computational redundancy is significantly reduced by decreasing the number of controllable modules and layers. Through this hardware-software co-design, 3D positional encoding mechanism, and text prompt word design, a single model can adaptively reconstruct high-precision images with multiple imaging modalities and compression rates.

The application discloses a hyperspectral compression imaging method, system, terminal and storage medium in the technical field of hyperspectral image processing, and aims to solve the problem of low reconstruction accuracy caused by high-frequency detail loss, weak spectral correlation and poor noise resistance in the compression imaging process in the prior art. It comprises obtaining the compression measurement and the perception mask of the coded aperture snapshot spectral imaging system, obtaining the initial feature map according to the compression measurement and the perception mask; inputting the initial feature map into the pre-constructed U-Net model to obtain the decoding feature map; and performing convolution mapping on the decoding feature map to obtain the reconstructed hyperspectral image in combination with the initial feature map; the spectral-space state synchronizer efficiently processes long sequence data, and the wavelet variance modulation block can effectively capture the local and global features of the image through multi-scale decomposition, thereby solving the problem of low reconstruction accuracy caused by high-frequency detail loss, weak spectral correlation and poor noise resistance in the traditional compression imaging process.

The application discloses a hyperspectral snapshot compressive imaging reconstruction and classification integrated method and system, which comprises a training stage and a test stage: in the training stage, a compressed measurement map is generated by using a hyperspectral image and coded aperture mask information, and a training sample composed of the compressed measurement map and a corresponding multi-label vector is constructed; the compressed measurement map is input into an end-to-end network composed of a shared feature encoder, an image reconstruction branch and a multi-scale fusion classification branch, and a joint loss function with dynamic weights is used to learn network parameters; in the test stage, the compressed measurement map collected by a snapshot compressive imaging system is input into the trained network, and a multi-label ground object classification result is directly output. Compared with a serial scheme of first reconstruction and then classification, the application reduces repeated calculation and parameter size, reduces inference delay, improves remote sensing interpretation efficiency, and is suitable for fast ground object identification scenes.

The application discloses a kind of training-free snapshot photometric stereo vision methods based on spectral compression imaging, comprising: (1) setting multiple different wavelength monochromatic light sources to simultaneously irradiate object from different directions, form high-dimensional spectral-angle data cube;(2) the spectral-angle data cube is modulated in space by coded aperture, along the spectrum dimension is sheared translation by using optical dispersion element, finally integrates on single-channel image sensor, generates single two-dimensional gray compressed measurement image;(3) inverse problem model is constructed, and multiple illumination image sequence is decoupled and recovered from single two-dimensional gray compressed measurement image using iterative optimization algorithm based on physical drive;(4) the surface normal map of the measured object is predicted by inputting multiple illumination image sequence into pre-trained photometric stereo vision network module.The application can realize low-cost, high data efficiency and high physical fidelity three-dimensional reconstruction by single snapshot.

A microscope system includes: a microscope acquiring an image of a specimen on a focal plane; an imaging unit imaging the image acquired by the microscope; a focal plane moving mechanism moving the focal plane in an optical axis direction of the microscope; a compression unit compressing image data pieces of the images acquired by the imaging unit to generate compressed image data; and a control unit transmitting, to the focal plane moving mechanism, a focal plane control signal to move the focal plane to a predetermined focal coordinate, the control unit transmitting, to the imaging unit, a plurality of imaging timing signals specifying the timing of imaging the images, the control unit transmitting, to the compression unit, layer count data indicating the number of the plurality of imaging timing signals.

The method of snapshot compressive imaging three-dimensional Gaussian splatting reconstruction belongs to the field of computer vision and computational photography. A reconstruction network based on the hybrid architecture of stochastic gradient Langevin dynamics (SGLD) and explicit three-dimensional Gaussian representation is built. Then, multiple physical perception modules are combined to achieve the deep decoupling and reconstruction of the spatiotemporal information of dynamic scenes. Driven by the embedded decoupling deformation field, the deformation field contains a coarse-grained trajectory prediction branch and a fine-grained integral fitting branch: the coarse-grained branch combines low-frequency time coding with latent embedding features, enabling Gaussian primitives to capture global rigid motion trajectories and serving as a temporal regularizer; the fine-grained branch uses high-frequency time coding to predict non-rigid deformation and physically fits the motion blur stripes generated by temporal integration through anisotropic scale stretching. An adaptive density control strategy compatible with physics is introduced. Using the physical integral imaging simulation module, high-fidelity and multi-view consistent high-speed dynamic three-dimensional scene reconstruction results are obtained.

The application discloses a kind of underground cable detection methods based on compression imaging technology, comprising: using ground penetrating radar to detect target area, obtain the echo signal of ground penetrating radar;Echo signal is pretreated;By compressed sensing method, the sparse representation of echo signal after pretreatment is analyzed, dictionary matrix is created, so that echo signal is converted from K·L matrix form to column vector form;Echo signal in column vector form is mapped to low-dimensional measurement vector y;According to measurement vector y, reconstruct target space b, and convert reconstructed target space b into K·L matrix, that is, target space imaging result.The application identifies underground cable by ground penetrating radar, wherein the application of compressed sensing algorithm can effectively reconstruct the cross-sectional image of underground condition, has good focusing effect, error is far less than before using under excavation method, and simultaneously the application of compressed sensing algorithm reduces the difficulty of reconstructed image, reduces the cost of underground cable detection, and detection efficiency is higher.

The invention discloses a super-resolution double-path time-space domain compression imaging method and device, and belongs to the field of high-speed high-resolution imaging and compression imaging. The system comprises an imaging objective lens, a beam splitter prism, an imaging lens, a high-spatial-resolution camera, a DMD, a low-speed low-spatial-resolution camera and a computer. The implementation method comprises the steps that a high-spatial-resolution imaging light path collects a single-frame high-spatial-resolution image, in a space-time coding compression measurement light path, target light is imaged to the surface of a micromirror array of a DMD through an imaging lens, primary space compression sampling of a target scene is completed at the DMD, and the target scene is subjected to spatial compression measurement; and meanwhile, the DMD controls the micromirror to turn over according to a pre-loaded time-space domain joint coding modulation matrix, so that time-space domain joint modulation of incident target light information is realized. And the low-speed low-spatial-resolution camera performs secondary spatial compression on the modulated multi-frame image, and performs integration on the multi-frame modulated image within single exposure time to obtain a single-frame low-spatial-resolution image. And the obtained high-spatial-resolution image and the space-time compressed image are solved through a reconstruction algorithm to reconstruct the high-spatial-temporal-resolution image.

The application provides a million-frame three-dimensional imaging device and method based on a high-speed rotating mirror, proposes a million-frame sinusoidal fringe generation module based on high-speed rotating mirror scanning, which can generate sinusoidal fringes at high speed, proposes a million-frame compressed imaging module based on high-speed rotating mirror scanning, and multiplexes a high-speed rotating mirror scanning reflection type random binary mask plate, so that the speed advantage of the high-speed rotating mirror is applied to the generation of sinusoidal fringes and the binary coding compression of high-speed images, and finally, the compressed images collected are subjected to three-dimensional image reconstruction by using a fast three-dimensional reconstruction algorithm, so that the million-frame three-dimensional imaging of structured light illumination is realized.

A snapshot compressed imaging reconstruction method and system, belonging to the field of image processing technology, solves the problems of insufficient intra-stage interaction and inter-stage information attenuation in existing technologies. The method includes: acquiring a two-dimensional compressed measurement image; performing initial inversion to obtain an initial reconstruction result; calculating a gradient feature map based on the previous stage reconstruction result and mask parameters, performing gradient descent updates on the previous stage reconstruction result to obtain the intermediate reconstruction result of the current stage; performing multi-scale processing on the gradient feature map to extract gradient features at different scales and generate weight information; reconstructing the intermediate reconstruction result of the current stage, using the weight information to modulate the feature interactions in the near-end mapping process to generate the reconstruction result of the current stage; passing the reconstruction result of the current stage to the next stage, establishing inter-stage message passing; S7: repeating the iteration until a preset number of stages is completed, and outputting the final reconstruction result. This invention is applicable to snapshot compressed imaging scenarios.

The embodiment of the application provides a kind of multi-modal single-pixel compression imaging system based on spatial light modulator, light source supply module in the system will provide wide band light source, the light emitted by wide band light source will be irradiated to target object, and reflected to light field modulation module;Light field modulation module will carry out optical spectrum separation to the received light signal, after the separated multiple single color light is incident to spatial light modulator, via spatial light modulator carries out light splitting and parallel modulation;Single-pixel detection module will when detecting modulated light signal, the light intensity data collected is fed back to data processing module;Data processing module will when determining that light signal reaches ideal modulation state, mode switching instruction is issued to spatial light modulator to carry out mode switching, and, based on multi-modal imaging light intensity data, joint optimization modeling and iterative solution calculation are carried out, to realize the synchronous reconstruction of target object image.The system can realize the joint acquisition and fusion reconstruction of multi-modal information.

A method of imaging a breast compressed with a paddle including a foam compressive element. The method emits an x-ray energy from an x-ray source towards the breast and the foam compressive element and a marker disposed adjacent the foam compressive element. The x-ray energy is emitted over a predetermined time period. The marker includes a physical characteristic. The x-ray energy is detected at a detector disposed opposite the breast from the x-ray source. An image of the compressed breast and the marker is generated based on the detected x-ray energy. The marker in the generated image includes an image characteristic associated with the physical characteristic. The image characteristic is compared to the physical characteristic. A signal is sent when the image characteristic deviates from the physical characteristic by a threshold.

The application discloses a snapshot spectral compression imaging method based on L1 norm and low rank technology, studies data fidelity and regularization in hyperspectral image reconstruction, models noise, and uses L1 norm to measure fidelity of the recovered hyperspectral image; a plug-and-play deep low rank prior is designed based on deep learning as a regularization term, the designed deep low rank prior is expressed as designing two neural networks to generate two matrices representing low rank characteristics of the hyperspectral image, random Gaussian noise is inputted to train the two networks in a self-supervised manner; the fidelity term with L1 norm and the regularization term with the deep low rank prior are combined as a hyperspectral image reconstruction optimization algorithm, and the hyperspectral image reconstruction optimization algorithm is solved based on a multiplier alternating direction method to obtain a target hyperspectral image. In the application of hyperspectral image reconstruction, the application has significant advantages, and higher fidelity and better imaging quality can be realized.