93 results about "Adaptive denoising" patented technology

The invention provides an intelligent event identification method and system based on a high-speed camera, and the method comprises the steps: setting the collection parameters of the high-speed camera, and triggering the camera to collect a target scene video stream. And hardware acceleration decoding processing is carried out on the collected original video data stream, real-time environment illumination information of the environment illumination sensor is obtained, and dynamic brightness equalization processing is executed. And performing motion adaptive denoising processing on the video sequence. Geometric distortion correction is carried out on the image sequence through camera calibration parameters, sub-pixel-level displacement vectors and dense optical flow field data of a moving object are extracted, and multi-scale morphological features are extracted. And the features are fused to generate motion feature data, the data are processed through a spatio-temporal joint event classification model, an event identification result is output, the result is bound with a high-precision timestamp, and event identification information is output to an industrial control system display device in real time. According to the invention, the accuracy and real-time performance of event identification can be improved.

The invention belongs to the technical field of artificial intelligence, and relates to an artificial intelligence-based electric energy quality disturbance identification and positioning method, which comprises the steps of constructing an electric energy quality disturbance signal data set, performing segmented preprocessing on electric energy quality disturbance voltage data, enhancing time-frequency joint features and encoding disturbance sensitive areas. And constructing a deep learning model for power quality disturbance identification and positioning, and identifying and positioning the power quality disturbance. According to the invention, through adaptive denoising processing, boundary detection and multi-resolution time-frequency feature extraction, the identification precision and positioning precision of power quality disturbance are significantly improved; self-adaptive wavelet denoising and dynamic segmentation are combined, noise interference is effectively suppressed, and the edge characteristics of voltage sudden change points are kept; according to the dual-task sharing network, disturbance identification and positioning tasks are cooperatively optimized, so that the network can consider disturbance classification and time positioning at the same time; and through Bayesian reasoning, the system can output confidence estimation, provides credibility quantification of identification and positioning results, and effectively improves the reliability of the system.

The invention discloses a multi-modal data adaptive denoising and missing reconstruction method and system, and relates to the technical field of point data denoising and reconstruction, and the method comprises the steps: obtaining to-be-processed multi-modal original data and a modal missing mask; performing unsupervised denoising on image data in the multi-modal original data to obtain a denoised image, and further obtaining multi-modal data; inputting the multi-modal data into a double-flow encoder for processing to obtain a multi-modal embedded vector of cross-modal alignment; the method comprises the following steps of: performing mapping and adding position embedding on a modal embedding vector to obtain each modal coding feature, determining a missing modal based on a modal missing mask, inputting an available modal coding feature into a retrieval enhanced expert model based on prototype memory to perform missing reconstruction to obtain a multi-modal joint representation, and mapping the multi-modal joint representation to a task output space through a full connection layer. Through introduction of unsupervised denoising, double-flow coding alignment and modal knowledge expert hybrid reconstruction, robust representation learning and information complementation under the condition that noise and modal missing exist in multi-modal data are realized.

The invention relates to the technical field of artificial intelligence image processing and biological detection crossing, and discloses a perception enhancement and feature distillation-oriented high-density microalgae detection method, which comprises the following steps: acquiring a high-density microalgae image containing cell overlapping, boundary blur and cross-scale distribution, and inputting the image into a backbone network to extract a multi-scale feature map; optimizing the bounding box through potential consistency mapping processing; generating a microalgae density map through density sensing auxiliary processing, and calculating density loss; self-adaptive denoising is carried out based on image complexity; a student model is optimized by using a dual-feature distillation framework, and the small-scale microalgae detection capability is enhanced; and finally, fusing the results of the modules, and outputting the position, category and quantity of the microalgae. According to the method, the overall average accuracy of high-density microalgae detection can be improved, bounding box jitter and small-scale microalgae omission ratio are reduced, the average reasoning time is shortened, the real-time detection requirement is met, and reliable data support is provided for microalgae culture process control and optimization.

The invention discloses a student psychological risk perception method based on multiple modes, and relates to the technical field of emotion calculation and intelligent education. The method comprises the following steps: firstly, extracting a facial expression feature vector and a voice intonation feature vector respectively by using a convolutional neural network and Fourier transform through a collected video stream and an audio stream; then adaptive denoising processing is carried out on environmental interference, timestamp alignment and dynamic time warping are carried out on the denoised multi-modal data, time sequence synchronization is ensured, and corrected multi-modal sequence data are formed; then, dynamic emotion track features are extracted from the sequence data, a preliminary emotion state label is generated by comparing the dynamic emotion track features with a baseline threshold value, and the threshold value is adaptively updated in combination with historical data so as to improve the judgment accuracy; and finally, aggregating the emotional state labels of a plurality of students to generate a visual group emotional thermodynamic diagram so as to realize macroscopic perception of group psychological risks. The accuracy, robustness and visualization degree of student psychological state analysis are effectively improved, and an efficient technical means is provided for campus psychological early warning.

The invention discloses an infrared visible light image fusion system based on semantic-driven space-frequency routing, and aims to solve the problems of multi-mode redundancy, weather degradation and insufficient space-frequency modulation. The core of the method comprises the following steps: a CMBA module realizes cross-modal bottom layer interaction and redundancy elimination; the FASRP module is integrated with 2D-SSM to perform long-range modeling, and high and low frequency feature decoupling is realized by using a routing mask; the DSPE module extracts environment degradation semantic priori by relying on a visual language large model; the DFMB module decouples the features into amplitude and phase, and realizes adaptive denoising according to prior dynamic gating. And the system reconstructs the image through fidelity fusion and progressive decoding. According to the method, the generalization ability in a complex scene is improved, and the fused image has precise infrared target positioning and visible light texture details.

The invention relates to the technical field of infrared imaging, in particular to an infrared image self-adaptive denoising algorithm and system based on an FPGA (Field Programmable Gate Array), and is characterized in that an adaptive blind pixel searching algorithm module captures data of a certain frame number at an FPGA end, performs multi-directional gradient detection on each pixel point, sets conditions to judge to obtain a blind pixel position, and marks the blind pixel position; the convolution kernel self-adaptive blind pixel replacement module performs self-adaptive blind pixel replacement multi-time module multiplexing at the PL end of the FPGA through a convolution kernel sliding block and gradient comparison to ensure complete coverage of a large blind pixel group, and the self-adaptive median filtering module intelligently selects and replaces pixels meeting requirements through automatic sorting of center pixels and surrounding pixels; according to the method, surface fuzzy filtering is carried out through a surface-blue algorithm module, so that a smooth area in an image is effectively smoothed, details of an edge area are reserved, blind pixels in a video stream can be effectively detected and compensated, and meanwhile traditional noise is suppressed and image details are reserved through multi-stage filtering.

The invention relates to a point cloud denoising method of a photon counting laser radar and an intelligent denoising system, and belongs to the technical field of photon laser radar data processing. The technical problems that the performance of a fixed parameter denoising method is seriously unbalanced under different ground feature and complexity scenes, and high denoising rate and high signal fidelity cannot be considered at the same time are solved. Synchronously acquiring an optical image and an original photon point cloud of the target area; identifying the dominant ground feature type of the laser spot area through real-time semantic segmentation; calculating scene complexity based on the point cloud data and determining the grade of the scene complexity; querying a preset two-dimensional parameter mapping table according to the combination of the surface feature type and the scene complexity level, and dynamically obtaining an optimal denoising parameter set; and finally, the parameter set is utilized to drive adaptive denoising processing, and signal points and noise points are distinguished by calculating the local density and direction consistency of the points and constructing a two-dimensional discrimination space. The method is mainly used for real-time processing platforms such as unmanned aerial vehicles, and the point cloud denoising precision, the signal retention rate and the topographic feature fidelity are remarkably improved.

The invention discloses a power transmission line unmanned aerial vehicle inspection depth visual intelligent identification method and system, and relates to the field of power transmission line inspection, and the method comprises the steps: dynamically adjusting imaging parameters based on the voltage grade and environment complexity of a power transmission line, collecting the multispectral visual data of a line body, a metal fitting for line connection and fixation, and a gallery environment, according to the invention, the method can dynamically adapt to the voltage grade of the power transmission line and the environmental complexity, optimizes the imaging effect, accurately collects the multi-dimensional visual data, and associates the key environmental parameters. The data standardization level is improved through adaptive noise reduction, accurate splicing and distortion correction, defect features and environment association attributes are efficiently mined through an attention guidance mechanism and a dynamic convolutional network, and accurate judgment of defect types and levels and future evolution trend prediction are performed in combination with historical data.

The invention relates to the technical field of graph contrast learning, in particular to a graph contrast learning recommendation method based on adaptive denoising enhancement and difficult negative sample generation. According to the method, a dynamic denoising mechanism is introduced to automatically and accurately recognize and filter noise edges, meanwhile, key information interaction in an original image is effectively reserved, linear mixing is carried out on embedding of an original view and a denoised view, then a loss function is optimized through two-stage joint, and the image quality is improved. The whole model is uniformly trained in combination with double-view representation and high-quality negative samples while noise is removed, and finally, the model can sensitively capture subtle differences between positive and negative samples through high-quality difficult negative samples generated adaptively, so that the discrimination ability of comparative learning is remarkably improved, and the accuracy of comparative learning is improved. The method not only strengthens the recognition capability of the model for the negative sample, but also optimizes the contrast learning target, effectively improves the overall learning effect and generalization capability, and provides a powerful guarantee for the accuracy of a personalized recommendation algorithm.

The invention provides a bidirectional LSTM photovoltaic power generation prediction method based on wavelet decomposition and double attention, and belongs to the technical field of power generation prediction. Selecting an optimal wavelet basis function by using a particle swarm optimization algorithm to carry out adaptive noise complete set empirical mode decomposition on the power sequence to obtain a multi-layer intrinsic mode function component, and carrying out adaptive denoising and power signal reconstruction according to a multi-scale permutation entropy and a Bayesian risk minimization criterion in combination with meteorological conditions; key feature variables are extracted through a maximum information coefficient, a bidirectional long-short-term memory network prediction framework is established, a feature attention mechanism and a double-path time attention structure are introduced, and when power mutation or irradiance mutation is detected, a sparse attention weight rapid reconstruction mechanism is triggered to complete prediction. The technical problem that the prediction precision is reduced when the photovoltaic power generation power changes suddenly under the cloudy weather condition is solved.

The invention provides a hyperspectral data optimization method and device, equipment and a storage medium, and relates to the technical field of image processing. The method comprises the following steps: acquiring to-be-processed hyperspectral data, and performing multi-scale feature extraction on the to-be-processed hyperspectral data to obtain a multi-scale feature vector; calculating the fuzzy correlation degree between different wavebands of the to-be-processed hyperspectral data, performing waveband screening according to the fuzzy correlation degree to obtain a reserved waveband, and performing dimensionality reduction on the to-be-processed hyperspectral data according to the reserved waveband to obtain dimensionality-reduced hyperspectral data; dividing the dimension-reduced hyperspectral data into a plurality of local areas, and performing adaptive denoising on each local area according to local data features of each local area to obtain denoised hyperspectral data; and performing data reconstruction according to the multi-scale feature vector and the denoised hyperspectral data to obtain optimized hyperspectral data. According to the method, the hyperspectral data can be optimized more efficiently and accurately, and the quality and the application value of the hyperspectral data are improved.

The invention discloses a self-adaptive noise reduction method for a vibration signal of a filling pipeline in the technical field of signal denoising. The self-adaptive noise reduction method comprises the following steps: decomposing the vibration signal of an original pipeline by adopting an optimized VMD to obtain a plurality of IMF components; identifying a signal dominant component and a noise dominant component in the IMF component; noise reduction is carried out on the noise dominant component by adopting an improved wavelet threshold function; and reconstructing the noise dominant component and the signal dominant component after noise reduction to obtain a denoised filling pipeline vibration signal. According to the method, the optimal modal number K and the penalty factor alpha of the VMD are adaptively determined by adopting an NSGA-II multi-target optimization algorithm, the subjectivity of parameter adjustment by traditional artificial experience is overcome, and meanwhile, the problems that a traditional single-target optimization method is prone to falling into local optimum, weak in global search capability and one-sided in optimization are solved; and meanwhile, a wavelet threshold function is cooperatively improved, so that medium-high-frequency noise can be eliminated, low-frequency interference can be suppressed, and the signal quality and the feature identification degree are greatly improved.

The invention discloses a method and a system for measuring the layer thickness of a multilayer hollow pipe fitting based on image recognition, and relates to the technical field of multilayer hollow pipe fitting detection, and the technical key points are as follows: firstly, enhancing boundary features through multi-band image fusion preprocessing, and offsetting dimensional deviation by combining dynamic environment compensation and residual stress correction; through multi-dimensional verification and error correction, the measurement precision reaches a preset high-precision standard and is obviously superior to that of traditional vernier caliper measurement and ultrasonic measurement; secondly, a stable chromatic aberration boundary is constructed through a color master adaptation rule, and the influence of material pores, defects and distribution uniformity is avoided; meanwhile, through the technologies of self-adaptive noise reduction, dynamic measurement density adjustment and the like, the influence of factors such as ambient light interference and boundary complexity is resisted, and the measurement reliability is ensured; in addition, the color masterbatch adaptation rule can be dynamically adjusted according to optical characteristics and forming process parameters of different materials, the method is suitable for various complex multilayer hollow pipe fittings, and a measurement scheme does not need to be redesigned for specific pipe fittings.

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.

The invention provides a self-adaptive denoising method based on Meyer wavelets and an improved smooth threshold function in order to solve the problem that a transformer partial discharge high-frequency current sensor field calibration signal is seriously polluted by field electromagnetic noise interference. The method comprises the following steps: firstly, arranging a standard sensor according to an optimal cross-core distance to obtain an original noisy signal; then adaptively determining a decomposition series according to a data length, selecting Meyer wavelets to perform multi-scale decomposition, performing smooth contraction on wavelet coefficients by adopting a continuously derivable, progressive and unbiased improved threshold function, and finally reconstructing a high signal-to-noise ratio verification signal through wavelet inverse transformation; the method does not need manual intervention of the threshold value, keeps orthogonality and high-order derivability in the whole process, and is beneficial to improving the accuracy of the verification result of the high-frequency sensor to be tested.

The application belongs to the field of intelligent ticket management, and specifically relates to a financial ticket information automatic input system based on artificial intelligence, which comprises a ticket collection module, a multi-scale self-adaptive noise reduction module, an intelligent classification module and an intelligent partition input module; the application adopts a feature double-domain self-adaptive enhancement method based on transpose self-attention, improves noise reduction robustness, specifically processes different noise types, overcomes the limitation of a single domain, restores the global structure while retaining the details, increases the noise reduction capability of the ticket image, and improves the text detection efficiency; the application adopts a document intelligent partition method, unifies heterogeneous tasks into the combination of instance and semantic segmentation, avoids model redundancy, maps the category name into a semantic query, supports open set classification and zero sample migration, realizes the dynamic interaction of instance and semantic query through hybrid query, enhances the understanding of the model to the complex document structure, and realizes efficient ticket image partition information extraction and input.

The application discloses a kind of based on adaptive noise reduction tool damage gray image intelligent classification method, mainly including a network consisting of adaptive noise reduction module and a tool damage image classification module;Noise reduction module and classification module are simultaneously accepted end-to-end joint training, share and optimize network parameters, while increasing a balance parameter in loss function, the parameter is according to the classification result of feedback of back propagation algorithm adaptive optimization noise reduction level, finally reach the optimal classification performance.The application can automatically identify whether the image contains noise and batch solve the classification problem of image noise for tool damage gray image, improve the prediction ability of model for difficult samples through adaptive noise reduction mode, can effectively remove image noise, while reducing the amplification effect of corresponding loss function to anti-noise, finally improve the intelligent classification ability of tool damage gray image from image processing quantity, image quality and prediction efficiency.

The invention relates to an intelligent aiming design method and system based on deep learning. According to the method and system, infrared and visible light bimodal video streams are synchronously collected, and weather types and grades are recognized in real time in combination with meteorological sensor data; carrying out image enhancement by adopting a physical constraint self-adaptive denoising network; constructing a target trajectory by using a target detection and tracking algorithm; high-precision prediction of a target motion track is realized based on a double-flow neural network and a time sequence prediction model; fusing the real-time meteorological data and the trajectory model to carry out trajectory compensation calculation; finally, the dynamic aiming point coordinates are calculated; according to the method, the technical problems of serious image degradation and misalignment of target prediction in severe weather are effectively solved, and the first hit rate and task efficiency of an aiming system in a complex environment are remarkably improved.

The application discloses a molecular property prediction method based on spectral position coding and bionic side inhibition gating, and relates to the field of olfactory perception. The method comprises the following steps: constructing a molecular structured input comprising atomic types, a Coulomb matrix and graph Laplacian eigenvector; fusing atomic chemical attributes and spectral domain topological positions through a feature extraction module to generate initial atomic features; using a multi-scale aggregation module, hierarchical neighborhood aggregation is carried out based on preset scale constraints to extract multi-level structure features covering chemical bonds, functional groups and molecular skeletons; with the help of a global interaction module of bionic side inhibition gating, local features are dynamically modulated by a global query signal to realize adaptive denoising and semantic sharpening; finally, through mask pooling and a decoupled multi-label prediction head, the scores of each odor attribute of the molecule are output. The application significantly improves the accuracy, structure perception ability and cross-task generalization performance of molecular property prediction.

The invention discloses a low-altitude target intelligent identification method and system based on large model fusion, and relates to the field of image identification, and the method comprises a capturing module which is used for shooting preset multi-spectral image data of an airplane takeoff and landing low-altitude flight region, and synchronously capturing the environment illumination, atmospheric visibility and fog concentration parameters of the airplane takeoff and landing low-altitude flight region; the preprocessing module is used for receiving the multispectral image and the environmental parameters shot and captured by the capturing module, and forwarding the multispectral image and the environmental parameters to the extraction module after performing registration fusion and adaptive denoising enhancement processing; the method combines multispectral image accurate registration fusion and adaptive denoising enhancement, effectively improves the image quality in a complex environment, deeply excavates and strengthens target key features and improves the feature discrimination by means of double-attention fusion and multi-scale feature level interaction, and improves the image quality through multi-dimensional feature matching and accurate coordinate solution reasoning. And target category rapid identification and space coordinate accurate locking are realized.

The application discloses a kind of pre-processing method and system for rock fissure infrared image analysis, it is related to infrared image processing technical field, the method includes: the noise level estimation is carried out to the original infrared image of rock fissure obtained, generates noise feature description information;Based on noise feature description information, the original infrared image is carried out adaptive denoising processing, generates denoising image;Based on denoising image, background separation and high-frequency component extraction are carried out, and generate high-frequency feature image;Based on high-frequency feature image, carry out detail enhancement processing, and generate detail enhancement image;Based on detail enhancement image, carry out adaptive brightness correction processing, and generate pre-processing output image.Adaptive noise suppression and edge preservation to rock fissure infrared image are realized, fissure high-frequency detail feature and trend continuity are strengthened, and the contrast of low brightness area is improved, and high-quality data basis is established for downstream detection and segmentation task.

The present application relates to the technical field of infrared imaging, and relates to an infrared image adaptive denoising method and system based on FPGA, which is suitable for finding a blind element algorithm module to capture a certain number of frames of data at the FPGA end, performing multi-directional gradient detection on each pixel point, setting a condition to determine the position of the blind element and marking it; a convolution kernel adaptive replacement blind element module is used to compare the convolution kernel slider and the gradient at the PL end of the FPGA, multiple reuse of the adaptive replacement blind element module is used to ensure complete coverage of the large blind element group, and the adaptive median filter module is used to intelligently select and replace the pixels that meet the requirements by automatically sorting the center pixel and the surrounding pixels; surface blur filtering is performed through a surface-blur algorithm module, so that the smooth areas in the image are effectively smoothed, while the details of the edge areas are retained, blind elements in the video stream can be effectively detected and compensated, and traditional noise can be inhibited through multi-stage filtering while the image details are retained.

The invention relates to the technical field of crossing of artificial intelligence and machine vision, discloses an AI-vision-based machine tool frame casting defect detection method and system, and aims to solve the technical problems that the image quality is reduced and the defect detection accuracy is influenced due to sudden change of instantaneous illumination in an industrial field. According to the method, illumination time sequence data are collected in real time by integrating a main imaging sensor and a high-frequency auxiliary light field sensor array, and future light field distribution is predicted based on a hardware Kalman filter; and generating a pixel-level dynamic gain matrix according to the main image, performing regional adaptive exposure compensation on the main image in a simulation domain, performing linear reconstruction and gain adaptive denoising, and outputting an illumination balanced image for an AI model to detect microcracks and pores. The system realizes microsecond-level illumination abrupt change response and local accurate compensation, and obviously improves the imaging consistency and defect detection rate.

The application provides a multi-modal high-definition camera fusion intelligent driving image re-projection error noise reduction method, comprising the following steps: S1: multi-modal camera joint space-time calibration and error reference library construction; S2: multi-modal image synchronous acquisition and standardized pretreatment; S3: driving scene semantic partitioning and error sensitivity level mapping; S4: bidirectional re-projection error tracing and attribution label generation; S5: pixel-level noise reduction weight adaptive allocation; S6: multi-modal fusion constraint iterative noise reduction and local closed loop verification; S7: error reference library self-updating and global parameter closed loop optimization. The application realizes high-precision re-projection error tracing and adaptive noise reduction, effectively suppresses errors while completely retaining the details of intelligent driving core perception targets, and improves the multi-modal image fusion quality and the reliability of the intelligent driving perception system.

The invention belongs to the technical field of medical image processing, and particularly relates to a functional magnetic resonance image data denoising method, system and device and a medium, and the method comprises the steps: obtaining child brain functional magnetic resonance original time sequence image data, child brain development prior map data, historical same-age denoising standard data and physiological movement synchronous collection data; brain tissue region segmentation and region-of-interest positioning are completed, positioning results of different brain regions are obtained, and time sequence signal features and space structure features of the different brain regions are extracted; constructing a children brain functional magnetic resonance multi-dimensional adaptive denoising model, generating a brain region noise type thermodynamic diagram, and generating a brain region hierarchical differential denoising strategy through a signal-noise separation algorithm to obtain denoised functional magnetic resonance image data; and generating a de-noising quality evaluation report and updating the parameter weight of the multi-dimensional adaptive de-noising model. Therefore, the problems of low denoising accuracy and the like in the prior art are solved.

The invention belongs to the technical field of power system fault detection and positioning. The invention discloses a power distribution network grounding fault accurate positioning method based on adaptive noise suppression and dual-target optimization. The method is characterized by comprising the following steps: 1) multi-modal feature extraction and adaptive denoising; 2) intelligently dividing a path solution domain based on a graph neural network; 3) double-target accurate positioning based on an improved multi-target grey wolf algorithm: constructing a double-target fitness function for the candidate section, wherein the first target is a path length error from a fault point to a main power supply, and the second target is a zero-sequence voltage phase matching degree; a self-adaptive weight factor is introduced, double targets are synchronously optimized by improving a grey wolf algorithm, the search range is expanded by combining a spiral search strategy, local optimum is avoided, and the accurate position of a fault is obtained through rapid solution. According to the method, the fault can be quickly and accurately positioned, and the operation reliability of the power distribution network is improved.

The invention relates to the technical field of intelligent sensors, in particular to an insulator defect evaluation method based on vibration acoustics, and the method comprises the steps: analyzing the amplitude difference between each non-backbone component and an acoustic signal, and the discrete degree of all amplitudes in each non-backbone component, determining the noise influence degree according to the number difference between all the non-backbone components and all the backbone components; determining a signal characteristic value based on the proportion of the average frequency spectrum amplitude of each backbone component in the average frequency spectrum amplitude of the acoustic signal and the random fluctuation degree of all the amplitudes in each backbone component, and determining the disturbance degree in combination with the sound influence degree; and based on the confusion degree and the disturbance degree of the acoustic signal, denoising the acoustic signal by adopting a wavelet function so as to evaluate the defect of the insulator. According to the invention, through adaptive denoising, environmental noise interference is suppressed, and the accuracy and reliability of insulator defect evaluation are improved.

The application discloses an infrared and visible light image fusion system based on semantic driving space-frequency routing, aiming at solving the problems of multi-modal redundancy, weather degradation and insufficient space-frequency modulation. The core comprises: a CMBA module realizes cross-modal bottom interaction and de-redundancy; a FASRP module integrates 2D-SSM for long-range modeling, and uses routing mask to realize high and low frequency feature decoupling; a DSPE module extracts environment degradation semantic priori based on a visual language large model; a DFMB module decouples the features into amplitude and phase, and realizes adaptive denoising according to the priori dynamic gating. The system finally reconstructs the image through fidelity fusion and progressive decoding. The application improves the generalization ability in complex scenes, and makes the fused image have both accurate infrared target positioning and visible light texture details.

The invention belongs to the technical field of computer vision, and particularly relates to a video image recognition system based on computer informatization. The data acquisition module is used for acquiring field environment data; the environment self-adaptive noise reduction module preprocesses the collected field environment data to obtain multi-modal data; the feature recognition module is used for processing the multi-modal data and outputting a species recognition result; the multi-dimensional spatio-temporal evolution graph construction module constructs a multi-dimensional spatio-temporal evolution graph according to the collected field environment data and the species identification result; the confidence evaluation and incremental learning module is used for calculating the confidence of the species recognition result according to the species recognition result and the field environment data, and starting small sample learning according to the species recognition result to update model parameters in the feature recognition module; the interpretable ecological report generation module is used for generating an interpretable ecological report according to the confidence coefficient, the multi-dimensional spatio-temporal evolution graph and the species identification result; the method improves the recognition accuracy, and has a good application prospect.