49 results about "Spatial response" patented technology

Methods and devices for analyzing a sample from a subject are provided. A device includes one or more coherent light sources and one or more partially coherent light sources. The light sources are configured to emit first and second wavelengths of electromagnetic radiation towards an imaging chamber configured to hold blood cells. The light sources are aligned to illuminate a single common area of the imaging chamber. The device includes optics positioned to receive light from the imaging chamber, and a detector in optical communication with the optics. The detector is configured to detect fluorescence emission and backscatter from blood cells when present in the imaging chamber. The device includes mixers / unmixers configured to replicate and mix first and second signals representative of the responses to the first and second wavelengths after interacting with the plurality of blood cells to generate spectro-spatial responses.

The invention discloses a gas recognition method based on multi-source information fusion and environmental perception, and the method comprises the steps: constructing a deep feature learning framework of multi-source fusion through combining the spatial response features, time sequence features and external environmental information of gas; the method comprises the following specific steps: preprocessing collected gas data, and respectively extracting features of an image mode, a sequence mode and an environment mode; fusing the image features and the sequence features through a cross attention fusion module, and capturing the space-time correlation of the data; a cross-modal attention compensation module is introduced, so that main-modal gas data adaptively gathers key information in an auxiliary-modal environment, and effective compensation of environmental factors on gas recognition performance is realized; and finally, gas category prediction is performed through a classification decision head. According to the method, the problems that the detection is easily interfered by environmental factors, the stability is poor or the qualification is inaccurate due to the fact that modeling depends on single modal data in the existing gas identification technology are solved.

The invention relates to the technical field of fault diagnosis, in particular to a fault diagnosis platform for parts in a cone box, which comprises a signal acquisition module, an abrupt change identification module, a space positioning module, a frequency band partitioning module and a grading judgment module. Original vibration signals in an operation period are collected, and signal data are divided according to a time sequence; according to the method, vibration signals are segmented according to a time sequence, energy feature induction is carried out, and dynamic comparison of continuous cycle parameter changes and space responses is combined, so that comprehensive extraction of multi-level features can be realized under data synchronization, energy zoning collection and correlation of multi-element information such as noise and temperature are realized, and a fault identification basis of multi-source parameter coupling is formed; fault characteristics are classified, abnormal grades are distinguished, a diagnosis platform is endowed with comprehensive discrimination capability for hidden anomalies, complex responses and grade distinguishing, and fusion monitoring and grading presentation of multi-dimensional data in a cone box operation state are promoted.

The invention provides a method for cooperatively regulating and controlling chlorophyll-a concentration variability through vortex-upwelling, and relates to the technical field of ocean remote sensing and environmental monitoring. A multi-source satellite and reanalysis data set is obtained, and a stationary time sequence is constructed; identifying a dominant variation period in chlorophyll-a concentration, sea surface temperature and sea surface height anomalies, and extracting a monthly abnormal value sequence; performing complex empirical orthogonal function analysis on the sea surface height anomaly, extracting a dominant spatial variation mode of the sea surface height anomaly, and identifying a vortex distribution position and form; classifying and screening target vortexes according to vortex types and seasons, carrying out normalization processing on the chlorophyll-a concentration, aligning to vortex centers, and analyzing a spatial response relationship between the vortexes and the chlorophyll-a concentration; and drawing a density time-depth profile map and a CHL-SLA composite map by combining surface chlorophyll-a concentration, sea surface height anomaly and vertical Argo buoy data, and analyzing an influence mechanism of vortex-upwelling on chlorophyll-a horizontal distribution.

The invention provides an underwater target detection system and method for an AUV-carried synthetic aperture sonar, and belongs to the technical field of underwater acoustic detection. The method comprises the following steps: transmitting an ultra-wideband sound wave signal to an underwater target, and collecting an echo signal by using receiving array elements which are non-uniformly and sparsely arranged along a track direction; preprocessing the echoes and decomposing the echoes into a plurality of non-overlapping sub-band signals, obtaining an equivalent space response by calculating inter-channel phase differences under each sub-band, and further constructing a joint observation matrix; and high-resolution azimuth scattering intensity distribution is recovered in combination with compressed sensing, and motion compensation is performed by using platform motion data to realize underwater imaging. According to the invention, the dependence of the traditional synthetic aperture technology on long-distance uniform-speed motion is broken through, and high-resolution and high-precision underwater target detection and imaging can still be realized under the condition of high-speed navigation of an AUV (Autonomous Underwater Vehicle).

The invention discloses a metal plate machining and forming method based on strain trajectory optimization, and particularly relates to the field of metal plate trajectory correction and complex control based on electrical variable and magnetic variable response analysis. And configuring the target position and number of clamping heads to enable the metal plate to finish initial clamping and positioning in the preset direction. Spatial response characteristics of an electrical variable and a magnetic variable are collected and analyzed in real time under the action of a multi-clamping structure, a metal plate track state recognition and pseudo-track correction mechanism is constructed, and dynamic coupling optimization of a pressing path and an actual strain state is achieved. The problems that clamping out-of-control areas are difficult to recognize and track states are difficult to judge in the dynamic adjustment condition of a clamping structure are solved, it is ensured that strain tracks which are closed accurately and consistent in shape face are formed in the continuous forming processes such as metal plate overturning, straightening and reshaping, and the forming quality and the process stability are improved.

Methods and devices for analyzing a sample from a subject are provided. A device includes one or more coherent light sources and one or more partially coherent light sources. The light sources are configured to emit first and second wavelengths of electromagnetic radiation towards an imaging chamber configured to hold blood cells. The light sources are aligned to illuminate a single common area of the imaging chamber. The device includes optics positioned to receive light from the imaging chamber, and a detector in optical communication with the optics. The detector is configured to detect fluorescence emission and backscatter from blood cells when present in the imaging chamber. The device includes mixers / unmixers configured to replicate and mix first and second signals representative of the responses to the first and second wavelengths after interacting with the plurality of blood cells to generate spectro-spatial responses.

The invention discloses a power transmission line monitoring and early warning system and method based on distributed optical fiber sensing, and relates to the technical field of power transmission line state monitoring, and the method comprises the following steps: S1, building a distributed optical fiber sensing transmission channel and a segmented reference coordinate system; s2, performing temperature-strain coupling frequency shift feature extraction based on the reference coordinate mapping matrix; s3, time sequence disturbance intensity inversion is carried out according to the coupling frequency shift characteristics; s4, performing adjacent section disturbance coupling rate analysis based on the disturbance intensity evolution sequence; s5, performing disturbance source position inversion and classification identification according to the coupling disturbance structure; s6, performing dynamic threshold early warning by using a disturbance source distribution structure; by setting a disturbance evolution tensor model and a spatial response function, system modeling of disturbance propagation paths, energy evolution and nonlinear interaction characteristics in a power transmission line is realized, the limitation of traditional analysis based on a single-point threshold value is broken through, and the positioning precision and the response analysis capability in a complex disturbance scene are effectively improved.

The invention belongs to the technical field of seismic exploration data processing, and discloses a seismic data reconstruction method and system based on dynamic region perception. According to the method, a seismic data reconstruction model is established, a dynamic region sensing basic module is introduced into a model encoder, a known original mask is utilized to carry out hard division on a feature map, independent normalization and affine transformation are respectively carried out on a complete region and a missing region, pollution of invalid information of the missing region to feature distribution of the complete region is blocked, and the accuracy of feature distribution is improved. And the precision and reliability of data reconstruction are improved. A dynamic region perception learning module is used in a model decoder, a learnable double-path pooling and convolution operation is used for dynamically generating a spatial response diagram with geological semantics, self-adaptive region division is achieved, the network can capture missing feature distribution which is left after multi-layer convolution and has fuzzy boundaries in a deep layer, and therefore the spatial response diagram with geological semantics is obtained. The problem of fuzzy mask division in a deep network is solved, and finer and more robust reconstruction of complex geological structures and weak signals is realized.

The invention relates to the technical field of partial discharge detection, in particular to a partial discharge detection system and method based on acoustic imaging. The partial discharge detection system comprises a laser excitation module, a distributed acoustic emission sensor array and a core control module. The laser excitation module emits a laser signal to the sensor array; the sensor arrays are installed on the single face of a tested device at equal intervals and collect partial discharge acoustic emission signals after receiving laser signals. The core control module constructs a time difference positioning constraint of the position of the partial discharge acoustic emission source, calculates a spatial response power function value corresponding to each spatial grid point coordinate in the tested equipment, and obtains a coordinate corresponding to a spatial response power function peak value as a candidate position coordinate; and verifying whether the coordinates of the candidate position conform to the time difference positioning constraint, and if so, taking the candidate position as the position of the partial discharge acoustic emission source. According to the invention, accurate detection of partial discharge is realized.

The application discloses a kind of based on natural language's traffic event's question and answer method and system, belong to intelligent question and answer field, first acquisition roadside image and user query, respectively by visual encoder and natural language encoder extraction multiscale visual feature set and text feature;Then utilize pre-training cross-modal intent alignment and focus module, with text feature as guide to spatial attention anchor and cross-modal intermingling to visual feature, output joint representation feature;Again, the feature is input light-weight decoder obtained by structured logic decoupling and migration strategy distillation, so that it executes step-by-step logical inference and coordinate regression based on spatial response graph in parallel, synchronously output natural language answer text and target bounding box coordinates;Finally generate traffic event question and answer result containing semantic answer and visual mark.The application effectively solves the problem that prior art cannot accurately and efficiently based on natural language to traffic event intelligent question and answer.

This invention relates to the field of defect detection technology, specifically to a high-voltage cable defect location system and method based on a two-dimensional ultrasonic array. The system includes a signal main peak acquisition module, a lateral deviation calculation module, an excitation timing optimization module, a main peak spatial focusing module, and a focusing area control module. This invention utilizes a multi-element spatial collaborative acquisition method of a two-dimensional ultrasonic array. It constructs a signal spatial mapping relationship through the main peak timing characteristics, combines path difference parameters to achieve fine analysis of the propagation characteristics of multi-source signals, optimizes the excitation timing and array spatial response structure by adjusting the excitation sequence, periodically determines spatial convergence indices, identifies the main peak focusing state of the defect area, and dynamically adjusts the element weight distribution to improve the signal response capability of specific spatial areas. The spatial location of the defect area is thus clearly distinguished, effectively improving the resolution and accuracy of multi-point defect location in complex cable structures.

The embodiment of the invention discloses an infrared band multi-mode calibration method and device. The method comprises the following steps: calculating a point spread function (PSF) of an original infrared image, and generating a spatial response matching filter function (SRMF) according to the PSF; carrying out two-dimensional convolution calculation on the SRMF to obtain an infrared image after effective space calibration; generating a possible spectral response function of ice pollution, and evaluating an optimal ice pollution spectral response through a cross calibration method; according to the optimal ice pollution spectral response, a spectral response function after possible spectral offset is generated, the optimal possible spectral offset is evaluated through a cross calibration method, and a spectral response function after spectral calibration is obtained through optimal sampling; calculating a linear coefficient, a nonlinear coefficient and a constant coefficient of radiometric calibration; and according to each coefficient of radiometric calibration, calculating the radiation quantity of the infrared image after multi-modal calibration. The method can overcome the existing problems and promote the quantification and multi-source fusion application of the satellite infrared data.

The application provides a spatial response compensation method and device, electronic equipment and a storage medium. The spatial response compensation method comprises the following steps: constructing a spatial transfer function according to spatial input information and corresponding spatial response information of a system; splitting the spatial transfer function into a minimum phase system and an all-pass system; determining a first inverse system corresponding to the minimum phase system and a second inverse system corresponding to the all-pass system; constructing a compensation system according to a preset target frequency response, the first inverse system and the second inverse system; compensating the spatial transfer function through the compensation system to obtain a compensated spatial transfer function; and converting the spatial input information of the system into target spatial response information through the compensated spatial transfer function. The application decomposes an arbitrary stable and causal linear time-invariant system into a minimum phase system and an all-pass system with different properties, and solves the inverse systems of the two systems, thereby reducing the complexity of the solution.

The embodiment of the invention discloses an ion source online detection method and device, electronic equipment and a storage medium, and relates to the technical field of ion sources, and the method comprises the steps: receiving an ion beam emitted by an ion source, and obtaining a first current signal and a second current signal based on the ion beam measurement; the first current signal is measured by moving the Faraday cup along a horizontal axis, and the second current signal is measured by the Faraday cup under at least one bias voltage; performing inversion according to the first current signal and a space response function of the Faraday cup to obtain beam density distribution of the ion source; and / or, according to the second current signal and the spatial response function, performing inversion to obtain ion energy distribution of the ion source; the spatial response function is used for representing the probability that the ion beam is collected after entering the Faraday cup; and carrying out on-line detection on the ion source based on the beam density distribution and / or the ion energy distribution. The embodiment of the invention is suitable for carrying out online detection on the large-size ion source, the detection mode is simple, and the cost is relatively low.

The application discloses a movable array element anti-interference position beam joint method facing unknown direction, introduces movable array elements, and makes array geometry structure be treated as adjustable optimization variable in deployment stage. By changing equivalent array element spacing and pattern structure of the array, the separable degree of expected signal and interference signal incident direction on the spatial response is improved, and the destructive influence of strong interference on the array output when falling in the high side lobe / grating lobe direction of the array is reduced. In each candidate position, an adaptive beam forming process with a pilot sequence as a reference is adopted to directly generate a beam forming weight vector corresponding to the position, so that the mean square error of the receiving output in the pilot interval is minimized. The "movable array element position-performance index" is modeled as a target function to be optimized, and a global adaptive search method based on a probability model is adopted to select the next evaluation position in an iterative manner, instead of exhausting evaluation on all candidate positions.

The embodiments of the present disclosure disclose an infrared waveband multi-modal calibration method and device. The method comprises: calculating a point spread function (PSF) of an original infrared image, and generating a spatial response matching filter function (SRMF) according to the PSF; performing two-dimensional convolution calculation on the SRMF to obtain an infrared image after effective spatial calibration; generating a possible spectral response function affected by ice pollution, and evaluating an optimal ice pollution spectral response by a cross-calibration method; generating a possible spectral response function after possible spectral shift according to the optimal ice pollution spectral response, and evaluating an optimal possible spectral shift amount by the cross-calibration method, and obtaining a spectral response function after spectral calibration by optimal sampling; calculating a linear coefficient, a nonlinear coefficient and a constant coefficient of radiation calibration; and calculating an infrared image radiation after multi-modal calibration according to the coefficients of radiation calibration and. The method can overcome existing problems and promote quantitative satellite infrared data and multi-source fusion application.

This application discloses an intelligent identification and calibration method for contamination of infrared flame detectors in tunnel environments, relating to the field of intelligent control identification. It addresses the technical problems of existing technologies, such as the inability to identify the type and spatial distribution of optical window contamination, the reliance on a single calibration strategy, and the difficulty in maintaining a stable alarm distance in complex tunnel environments. The method includes: acquiring data within the current detection cycle; calculating the transmittance deviation between the transmittance characterization value and a pre-calibrated reference transmittance characterization value; acquiring spatial response deviation information; acquiring at least one type of tunnel operating condition information within the tunnel environment; determining the contamination mode of the optical window based on the transmittance deviation, spatial response deviation information, and tunnel operating condition information; selecting a calibration strategy matching the contamination mode of the optical window from a set of preset calibration strategies based on the contamination mode of the optical window; and using the selected calibration strategy to perform sensitivity compensation calibration on the infrared flame detection signal.

The application discloses a kind of hybrid codec parallel single-pixel imaging method, it is related to the field of computational imaging technology.The specific includes: configure multiple photoelectric detectors to receive light signal from target scene, and the target scene is regarded as by multiple sub-regions;Control spatial light modulator to project a series of encoding patterns to target scene;Utilize multiple photoelectric detectors and collect light intensity signal in parallel, wherein each photoelectric detector collected light intensity signal contains the mixed light intensity component from multiple sub-regions;Obtain a unmixing matrix, the unmixing matrix represents the spatial response weight distribution of each photoelectric detector to each sub-region light signal;Based on unmixing matrix, the light intensity signal collected is unmixing operation to separate out respectively corresponding to each sub-region independent response signal;According to independent response signal and encoding pattern, reconstruct the image of each sub-region.The purpose is to solve the problem of traditional single-pixel imaging slow and array detector high cost, there is signal crosstalk.

The present application relates to the technical field of partial discharge detection, and particularly relates to a partial discharge detection system and method based on acoustic imaging. The partial discharge detection system comprises a laser excitation module, a distributed acoustic emission sensor array and a core control module. The laser excitation module emits laser signals to the sensor array; the sensor array is installed at an equal interval on a single surface of the measured equipment, and collects partial discharge acoustic emission signals after receiving the laser signals. The core control module constructs a time difference positioning constraint of the partial discharge acoustic emission source position, calculates spatial response power function values corresponding to the coordinates of each spatial grid point in the measured equipment, obtains the coordinates corresponding to the peak value of the spatial response power function as candidate position coordinates; verifies whether the candidate position coordinates meet the time difference positioning constraint, and if so, the candidate position is taken as the position of the partial discharge acoustic emission source. The present application realizes accurate detection of partial discharge.

This invention provides a wireless communication control method and apparatus utilizing element error fitting of a millimeter-wave array antenna. The method includes: identifying anomalous element positions based on the synchronization relationship between the received signal sequences of each element obtained from the downlink reference signal received by the millimeter-wave array antenna during normal communication, thus obtaining anomalous elements; constructing local spatial response difference features based on the first received signal of the anomalous element in the received signal sequence and the second received signal of the normal element in the received signal sequence; fitting the offset of each anomalous element relative to the ideal element response based on the projection behavior of the local spatial response difference features in the spatial angular domain, thus obtaining the element error morphology; and activating redundant antenna subarrays if the element error morphology indicates that the millimeter-wave array antenna does not meet preset beam pointing conditions. This invention improves the robustness and stability of millimeter-wave communication systems under complex hardware conditions.

This invention relates to a fine-grained classification method and system for unbalanced clothing based on attention enhancement, belonging to the field of computer vision and image recognition technology. It extracts a first feature map and a second feature map to obtain a first spatial response map and a second spatial response map. Based on the pixel activation intensity distribution and pixel activation statistical characteristics of the response maps, it obtains an overlap rate index and an adaptive threshold. Spatial filtering is performed on the first feature map to obtain a filtered first feature map, and cross-level residual enhancement processing is performed on it to output a target discrimination feature vector. The classification result and its feedback optimization parameters are obtained to update the adaptive threshold. This method helps to eliminate the interference of background noise and spurious activation on fine-grained discrimination, significantly improving the robustness and practical value of the classification system.

The invention provides a large receptive field lightweight target detection method and system based on improved YOLOv11n, and relates to the technical field of computer vision and artificial intelligence. According to the large receptive field lightweight target detection method and system based on the improved YOLOv11n provided by the invention, a large kernel convolution structure (MarK Block + Dilated Reparam Block) is introduced into a YOLOv11n backbone network, and a spatial perception re-calibration mechanism (SRD) is combined, so that the receptive field and spatial response capability is remarkably enhanced, and the detection efficiency is improved. Long-distance dependence features such as decay, shielding and irregular textures in scenes such as detection of pigs died of illness can be more effectively captured, and the target sensing accuracy and robustness under a complex background are improved.

This invention relates to the field of aviation technology in image processing, specifically disclosing an aviation smoke detection method based on image recognition. The method acquires continuous image sequences from multiple perspectives within the cabin, constructs a four-dimensional spatiotemporal tensor, extracts temporal feature sequences along the time axis, and performs directional convolution sampling based on airflow direction parameters to obtain spatiotemporal response values, which are then aggregated into a feature map set. Spatial weights are generated based on the physical laws of smoke diffusion, and temporal weights are generated based on grayscale fluctuation frequency; these weighted values ​​yield a feature representation. The feature representation is decomposed into spatial response maps, and the three-dimensional coordinate parameters of the smoke source are obtained through spatial coordinate voting and weighted fusion. A confidence score is calculated based on the dispersion of candidate coordinates and airflow paths. The coordinate information is matched to cabin spatial layout data to generate a situational output. This invention can achieve precise location of smoke sources and output reliability quantification indicators in strong airflow environments.

The application discloses a chirp signal direction of arrival estimation method based on fractional domain deconvolution beam forming, and the method comprises the following steps: performing fractional order Fourier transform on array receiving signals by using optimal focusing order; generating fractional domain steering vectors according to array structure and optimal focusing order, and calculating beam forming output for all beam scanning directions; generating fractional domain array directivity patterns according to array structure and optimal focusing order, removing the influence of array spatial response by using deconvolution algorithm, and restoring real target spatial distribution; obtaining direction of arrival estimation results by peak searching based on the restored target spatial distribution; and realizing underwater detection and positioning based on the estimation results. The application realizes high-resolution direction of arrival estimation of a wideband chirp signal under the condition of a small aperture and low signal-to-noise ratio, and further realizes accurate detection and positioning of an underwater target.

The application provides an underwater target detection system and method for an AUV carrying a synthetic aperture sonar, and belongs to the technical field of underwater acoustic detection. The method transmits an ultra-wideband acoustic wave signal to an underwater target, and uses non-uniform sparse arrangement of receiving elements along the track direction to collect echo signals; the echo is preprocessed and decomposed into multiple non-overlapping sub-band signals, the equivalent spatial response is obtained by calculating the inter-channel phase difference in each sub-band, and then a joint observation matrix is constructed; high-resolution azimuth scattering intensity distribution is restored by combining compressed sensing, and motion compensation is performed using platform motion data to realize underwater imaging. The application breaks through the dependence of traditional synthetic aperture technology on long-distance uniform motion, and can still realize high-resolution, high-precision underwater target detection and imaging under the condition of high-speed navigation of the AUV.

This invention belongs to the field of seismic exploration data processing technology and discloses a seismic data reconstruction method and system based on dynamic region perception. The method constructs a seismic data reconstruction model, introduces a dynamic region perception module into the model encoder, uses a known original mask to hard-divide the feature map, and performs independent normalization and affine transformation on complete and missing regions respectively, preventing invalid information from missing regions from contaminating the feature distribution of complete regions, thus improving the accuracy and reliability of data reconstruction. In the model decoder, a dynamic region perception learning module is used to dynamically generate a spatial response map with geological semantics using learnable dual-path pooling and convolution operations, achieving adaptive region partitioning. This allows the network to capture the distribution of missing features with blurred boundaries remaining after multiple convolutions in deep layers, solving the problem of blurred mask partitioning in deep networks and achieving more refined and robust reconstruction of complex geological structures and weak signals.

The present invention discloses a metal plate processing and forming method based on strain trajectory optimization, specifically relating to the field of metal plate trajectory correction and reshaping control based on the response analysis of electrical and magnetic variables. The method includes placing the metal plate to be processed on a clamping platform and configuring the target position and number of clamping heads according to the size and shape parameters of the metal plate, so that the metal plate completes the initial clamping and positioning along a preset direction. By collecting and analyzing the spatial response characteristics of electrical and magnetic variables in real time under the action of multiple clamping structures, a metal plate trajectory state identification and pseudo-trajectory correction mechanism is constructed, and dynamic coupling optimization of the pressing path and the actual strain state is achieved. This solves the problem of difficulty in identifying the clamping out-of-control area and misjudging the trajectory state under the dynamic adjustment conditions of the clamping structure, ensuring that a closed, accurate, and consistent strain trajectory is formed during the continuous forming processes such as metal plate flipping, straightening, and reshaping, thereby improving the forming quality and process stability.

The invention provides a method for testing noise equivalent irradiance, which comprises the following steps of: acquiring multi-frame target image data obtained by imaging a surface source target by a tested infrared imaging device, the surface source target having a temperature different from that of a background; processing the multiple frames of target image data to obtain a filtered image, and based on the filtered image, calculating to obtain a signal-to-noise ratio between a target and a background and a radiation illumination difference at an entrance pupil; based on the signal-to-noise ratio and the irradiance difference, calculating to obtain intermediate noise equivalent irradiance; the diffusion characteristics of the target and background edges in the target image data are analyzed, and a diffusion compensation coefficient used for correcting the spatial response influence is obtained through calculation; and according to the intermediate noise equivalent irradiance and the dispersion compensation coefficient, calculating to obtain the target noise equivalent irradiance. According to the method, a brand new solution which is simple and convenient to operate, good in environmental adaptability, high in test precision, good in repeatability, high in precision and high in efficiency is provided for testing the NEFD parameters of the infrared imaging device.

The application discloses an underwater image sharpening method based on differential convolution and polarization fusion, comprising: spatially registering an RGB image and a polarization image of an underwater object; inputting the RGB image into a fusion differential convolution backbone network to cooperatively extract edge texture, color background and structure information of the image through differential convolution and conventional convolution, and obtaining embedded features; inputting the polarization image into a polarization coding branch to guide interactive fusion of the embedded features at different scales through multi-scale, and obtaining multi-modal fusion features; constructing an RGB spatial enhancement branch based on spatial distribution of the fusion features to enhance spatial response of a target region and significant texture; and obtaining a sharpened underwater image through a decoder and up-sampling reconstruction. The application effectively solves problems such as image detail blur, texture degradation and color shift in a high-turbidity underwater environment, enhances detail information and improves image sharpness while realizing color correction, and significantly improves underwater image quality.