82results about How to "Improve feature extraction" patented technology

The invention discloses a real-time infrared small target detection method based on a linear global scanning network. In order to solve the problems that an existing model is high in calculation complexity and poor in real-time performance, a U-Net-like lightweight architecture giving consideration to both detection precision and reasoning speed is constructed. In the encoding stage, shallow local textures are extracted through Stem and ResBlock, a linear global scanning (LGS) module is introduced into a deep layer, and the core of the LGS module captures anisotropic long-distance semantic dependency with linear complexity by utilizing space scanning GRU. The coding end multiplexes the GRU by using a linear context aggregator (LCA) and combines a channel reweighting enhancement feature. In the decoding stage, jump connection semantics are aligned through PlainBlock, and upsampling features are fused through a StandFusion module and are refined through cascade convolution. And finally, an Inception module is combined with depth supervision to generate a high-precision prediction result. The method is low in calculation overhead, high in detection precision and suitable for high-frame-rate real-time infrared monitoring scenes.

The invention discloses a distributed photovoltaic panel anomaly detection method under a high-altitude view angle based on a YOLOv8n model, and belongs to the technical field of target detection. The model comprises the following steps: acquiring a distributed photovoltaic panel image data set under a high-altitude view angle, and performing data enhancement and labeling; an improved YOLOv8n model is constructed, and the improvement comprises the steps that a C2fAT module is introduced into a backbone network, and the small target feature extraction capacity is enhanced; an SPPF module is replaced by an SPPF-LSKA module, and complex background interference is suppressed; an EMA attention mechanism is introduced into the neck network, and the multi-scale adaptive capacity is improved; optimizing a training process by adopting a WIoU v3 loss function; and training and optimizing the model by using the training set, and finally outputting a detection result through the test set. According to the invention, the problems of false detection and missing detection of the distributed photovoltaic panel in a high-altitude view angle are effectively solved, the detection precision of the distributed photovoltaic panel is further improved, and inspection personnel are helped to troubleshoot the photovoltaic panel in an abnormal state in the high-altitude view angle.

This invention discloses a bearing remaining life prediction method based on frequency domain degradation perception. It constructs an integrated prediction system through a time-frequency decomposition dual-branch feature extraction module, a frequency domain degradation perception weight allocation module, and a life prediction loss function optimization module. First, the time-frequency decomposition dual-branch feature extraction module decomposes the original time series into high and low frequencies and inputs it into an sLSTM and mLSTM dual-branch structure for feature extraction. Second, the frequency domain degradation perception weight allocation module performs degradation perception processing on the fused features, generating dynamic weights and modulating the features to obtain the final fused features, which are then mapped to the prediction space. Finally, the life prediction loss function optimization module calculates the loss, optimizes the model output, and obtains the bearing remaining life prediction result. This invention comprehensively covers the full-stage features of bearings from healthy to severely degraded, improves the sensitivity of early fault detection, and still possesses excellent generalization and robustness under complex operating conditions.

The application discloses a voltage sag identification method based on a triple tuple prototype network under a small sample, and belongs to the technical field of power quality analysis. The method uses a triple tuple feature extractor, a large number of voltage sag triple tuples are constructed, and effective sag features can be extracted by the model under the condition of a small number of training samples. Then, in view of the problem that some voltage sag features are similar and easy to confuse, an efficient channel attention mechanism is integrated into the triple tuple feature extractor, cross-channel feature interaction information is captured under the condition of only a small number of parameters, the model can pay attention to the key feature area, a prototype classifier is finally constructed, representative class prototypes are learned for each class by using the extracted sag features, and the final sample class is determined by comparing the similarity between sample features and class prototypes. Under the condition of limited sample data, the method can realize accurate voltage sag classification effect, and has strong practical application significance.

The invention discloses a PHSM-YOLO-based multi-target wearing identification method. The PHSM-YOLO-based multi-target wearing identification method comprises the following steps: constructing a PHSM-YOLO target detection model used for identifying whether a constructor correctly wears a safety appliance; collecting a target image containing constructors in the construction environment; and inputting the target image into a PHSM-YOLO target detection model, and outputting a wearing identification result of the safety wearing tool of each construction worker in the target image by the PHSM-YOLO target detection model. According to the multi-target wearing identification method, the wearing of the safety wearing tool can be identified on the target image containing the construction personnel through the constructed PHSM-YOLO target detection model, so that the personal safety of the construction personnel is guaranteed.

The invention relates to a living body detection method and device, an electronic device and a storage medium, and is applied to the field of target recognition, and the method comprises the steps: inputting a to-be-detected image collected by monocular equipment into a target lightweight network structure in a living body detection network, and obtaining a face detection frame and a living body label corresponding to the face detection frame; and the target lightweight network structure is used for acquiring target global features, including dense semantic information in the to-be-detected image, in the to-be-detected image. Cutting a human face detection frame corresponding to the target living body label containing the living body content to obtain a cut human face, and inputting the cut human face into the target texture discrimination network to obtain a living body detection result; and the target texture judgment network is used for carrying out classification judgment on the clipped face on a texture feature level. According to the method and the device, the problem that hardware cost saving and accurate face detection effect keeping cannot be taken into consideration under the condition that only one-way color image input exists in the prior art is solved.

The present application relates to artificial intelligence technology, and discloses a video feature extraction model training method, comprising: dividing a video sample into frames to obtain a first video frame set, performing video frame masking on the first video frame set to obtain a second video frame set, performing image block masking on image blocks of video frames that have not been subjected to video frame masking processing to obtain a third video frame set, performing video frame prediction training on a pre-constructed feature extraction model using the second video frame set and the third video frame set, calculating a loss value between a video frame prediction result and the first video frame set, adjusting parameters of the feature extraction model according to the loss value, and obtaining a trained video feature extraction model when the adjusted feature extraction model meets a preset model training condition. The present application also proposes a video feature extraction model training device, equipment and medium. The present application can improve the feature extraction capability of the video feature extraction model.

The invention provides a fan blade crack detection method, which comprises the following steps of: for a fan blade image acquired by an unmanned aerial vehicle, performing deblurring processing on the image through a pre-processing algorithm designed by the invention, and overcoming the problem of motion blurring when the unmanned aerial vehicle acquires the fan blade image; and carrying out crack detection on the acquired image by adopting a crack detection neural network model improved based on YOLOv8. According to the crack detection model, self texture interference of the blade can be avoided, the crack detection precision is improved, crack characteristics of different scales are captured through a Stem module of a multi-branch structure, a CAM module is added into a C2f-1 module, crack attention is integrated, and crack related characteristics are enhanced; a residual connection mode of a residual block is designed in a C2f-2 module to avoid the problem of deep network gradient disappearance, gradient flow and effective transmission of crack characteristics are guaranteed, response of crack-related channels is enhanced through channel re-calibration, and irrelevant channels such as blade textures are inhibited.

This invention discloses a target region segmentation method and system for spatiotemporal event pulse streams, belonging to the field of video segmentation. This invention effectively adapts to the input of traditional neural networks by transforming continuous event pulse streams into specific input representations; it improves the model's feature extraction capability for continuous event pulse inputs by extracting information and memorizing past features through recurrent neural networks; it improves the model's query matching capability by matching targets and updating hidden states simultaneously through a recurrent feature encoder to model spatiotemporal relationships; it improves the model's robustness to feature matching by obtaining the attention relationship matrix of the target at the current and past moments through a feature decoder; and it improves the target region segmentation model's feature fusion capability for coarse-grained and fine-grained information by inputting intermediate features into the segmentation head through skip connections, thereby improving the accuracy of predicting target regions. This invention is adaptable to low-light and high-speed scenes and can segment target regions at any time.

The present application relates to a kind of forest leaf instance segmentation method and system based on deep learning, method includes: obtaining vegetation image, vegetation image is input into leaf instance segmentation model, obtains leaf instance segmentation prediction result;Leaf instance segmentation model is trained using training set;Training set includes: vegetation original image;Feature extraction and enhancement are carried out using backbone module in leaf instance segmentation model, and adaptive spatial fusion mechanism in progressive feature pyramid network is integrated to dynamically adjust feature weight, generate dynamic fusion feature;Through the dynamic asymmetric spatial perception mechanism built-in in dynamic anomaly regression head module, the corresponding multi-source deformation feature layer of dynamic fusion feature is obtained, and the feature fusion strategy of top-down cascaded decoding module is used to optimize multi-scale feature, obtain multi-source fusion feature layer, further using multi-source fusion feature layer, generate leaf instance segmentation prediction result.The present application solves the problems of data scarcity, poor adaptability and low efficiency.

The invention provides a reinforcement learning sequential sequence analysis system and method based on deep hybrid experts, and the system comprises a preprocessing module which carries out the denoising of a sequential sequence, and carries out the division of the denoised sequential sequence according to a fixed window size, and obtains a preprocessed sequential sequence; the strategy network is used for coding and decoding the time sequence processed by the preprocessing module on the basis of a hybrid expert layer to obtain scoring information of the time sequence; and the weight generator is used for obtaining the weight of each time sequence based on the scoring information of the strategy network. Through the data preprocessing module, the strategy network and the weight generator, better sequential sequence feature extraction and better sequential sequence scoring are realized.

The present application relates to the technical field of medium and shallow coalbed methane development, in particular to a kind of medium and shallow coalbed methane well production dynamic prediction method, comprising the following steps: S1, target well data acquisition and preprocessing;S2, random forest algorithm filters production main control factor;S3, fusion attention mechanism's CNN-LSTM multimodal time series feature extraction;S4, physical constraint deep learning model construction;S5, mixed loss function construction and model training;S6, production dynamic prediction and result output.The present application fuses physical constraint and deep learning technology, both utilize the efficient feature extraction capability of deep learning, and also guarantee the physical consistency of prediction result by physical constraint, solve the problem that pure data driven model generalization ability is poor, long-term prediction error is big.

This application discloses a method for training a recognition model, a recognition method, a computer device, and a computer-readable storage medium. The training method includes the following steps: controlling a single-photon lidar to emit photon pulses towards a target scene and receiving one-dimensional single-photon echoes reflected from the target scene, where the target scene includes at least one object to be recognized; constructing a training sample set based on the one-dimensional single-photon echoes; constructing an initial model and inputting the training sample set into the initial model for iterative training; validating the initial model after each iteration, marking the validated initial model as the recognition model, and outputting it. The application method involves inputting the echo signal to be recognized into the trained recognition model, and obtaining the recognition result based on the processing of the recognition model. The recognition result includes the pose and type of the object in the target scene. Therefore, this application can achieve high-precision recognition of long-distance targets.

The application discloses an image semantic segmentation model which is composed of a designed mixed attention focusing method, an attention rectification residual module (ARRM) and a mixed feature integration module (MFFM), is trained in a deep supervision mode as a whole, is reasonably provided with deformable convolution, is combined with a constructed multi-scale spatial attention module (MSP) and a double-pooling attention module (DPA) to be jointly optimized, and the problem of small target feature segmentation difficulty is solved. In order to reflect the excellent performance of the model on a downstream task, the training between different data sets is completed in a transfer learning mode, and the application range of the model is expanded. Finally, grouping convolution is added in the backbone network, the calculation cost is greatly reduced, and the problem of high training cost of the segmentation model is solved under the premise of guaranteeing the segmentation effect.

The invention provides a cross-working-condition fault diagnosis method and system for a gearbox of a mining shovel loading robot. The method comprises the steps that S1, vibration signals of the gearbox of a lifting mechanism of the mining shovel loading robot under different working conditions are collected; s2, constructing a diagnosis network and performing adaptive feature extraction on the vibration signals; the diagnosis network comprises a feature extractor and a fault classifier of the one-dimensional deformable convolutional network; s3, training the diagnosis network by adopting a dual-stage domain adaptive strategy fused with triple loss, and outputting a diagnosis model; and S4, performing fault diagnosis on the vibration signal under the target working condition according to the diagnosis model, and outputting a diagnosis result. According to the method, a double-stage domain self-adaptive strategy is adopted, feature optimization is carried out on the source domain in the first stage, data enhancement and domain invariant feature learning are carried out through mixed source domain and target domain labeling samples in the second stage, the domain offset problem caused by working condition changes is effectively relieved, and cross-working-condition efficient fault diagnosis is achieved.

The invention discloses a ceramic membrane internal defect detection method based on parallel stripe convolution and topological convolution. The method comprises the following steps: 1) constructing a ceramic membrane plate internal defect data set; 2) in a feature extraction module, providing a parallel multi-branch strip convolution feature extraction module, and better extracting features while realizing light weight; in a detection head module, a self-adaptive topological convolution detection head is provided, and the module reduces calculation of redundant features; and 3) model training and optimization: carrying out multiple rounds of training on the self-built data set. According to the method, for the problem of YOLO model lightweight, parallel strip convolution is used in a feature extraction part, and features can be better extracted on the premise of reducing the network parameter quantity; in addition, adaptive topological convolution is used in a detection head part, so that redundant calculation of a large-area non-key background area can be effectively reduced.

The invention belongs to the related technical field of image denoising, and provides an image denoising method and system based on physical information guidance in order to solve the problems that an existing eye fundus image speckle noise suppression method is weak in generalization ability, limited in denoising and the like. Randomly fusing the registered average living eye fundus image and the false eye static average image into a feature map to train a speckle noise estimation network; in the training process, the estimation loss is calculated based on the prediction result of the speckle noise estimation network and the mask region feature of the false eye static average image of which the signal-to-noise ratio tends to be stable; and noise distribution perception loss is calculated based on the prediction result of the speckle noise estimation network and the feature maps corresponding to the false eye static average image with the signal-to-noise ratio tending to be stable, and then training of the speckle noise estimation network is guided. According to the method, a high-signal-to-noise-ratio false eye still image is used as physical information to guide a speckle noise estimation network to realize high-signal-to-noise-ratio denoising under different noise distributions.

The application belongs to the field of multi-target tracking. An anti-occlusion target tracking method fusing multi-granularity dynamic appearance is provided. The purpose is to solve the technical problems of target ID jump and tracking effect affected in the occlusion scene in the prior art. The main scheme includes obtaining an original image. A target detector is called on the original image to obtain a target detection result, and a target appearance feature is extracted from the target image in the target detection result; a target appearance contribution factor alpha of the video frame is calculated; based on the target detection result of the previous frame, a target motion feature of the current frame is obtained. The target appearance feature, the target appearance contribution factor alpha and the target motion feature are data fused to obtain a target similarity. The targets in two frames are cascade matched and IOU matched, and each target is assigned an ID. The positions of the targets associated with the same ID in each frame are obtained, the running track of the same target ID in the image sequence is obtained, and the target tracking result is output.

The invention relates to a non-ideal array DOA estimation method based on multi-scale weight distribution and transfer learning and electronic equipment, and the method comprises the steps: constructing a fine tuning data set under a non-ideal array condition, the fine tuning data set comprising array signal data simulating various array physical defects; constructing an improved denoising and classification network, and training the improved denoising and classification network by adopting a transfer learning strategy; inputting a test signal in a non-ideal array environment into the trained improved denoising and classification network, and performing DOA estimation; the electronic equipment is realized based on the method. According to the method, the calculation overhead is reduced, the model is more robust to the interference of array errors and noise, the feature expression ability of signals is improved, the robustness and precision of the model in a complex environment are improved, transfer learning enables the model to obtain better performance in a low signal-to-noise ratio environment through a parameter sharing mode, and the method is suitable for large-scale popularization and application. And new non-ideal array data can be quickly adapted, and the retraining time and computing resources are reduced.

The invention discloses a bearing fault identification method based on multi-scale space-time synchronization attention and space-time alignment. The method comprises the following steps: carrying out sliding window segmentation and normalization preprocessing on an original vibration signal; constructing an adjacent matrix by using a K-nearest neighbor algorithm, and extracting local spatial features by using a graph convolutional neural network; bidirectional sequence features are extracted through a bidirectional gating loop unit network, and weighted fusion is carried out in combination with a global attention mechanism; and after space and time sequence features are spliced, fault prediction is realized through adaptive average pooling and a full-connection classifier. According to the method, graph modeling, GCN, BiGRU and an attention mechanism are fused, cooperative extraction of space-time double-path features is realized, the problems of weak space modeling and lack of time sequence dependence in a traditional method are effectively solved, and the bearing fault recognition precision and generalization ability in a multi-working-condition and strong-noise environment are remarkably improved.

The application belongs to the technical field of tunnel lining detection, and specifically discloses a ground penetrating radar high signal-to-noise ratio image imaging method, which comprises the following steps: constructing a training data set of pairs of noisy ground penetrating radar B-scan images and clean images; building an improved U-Net denoising network model, embedding a CBAM attention module in the encoder and the decoder, and enhancing effective signal features and suppressing noise through channel attention and spatial attention; using the training data set to train the model, optimizing network parameters with clean images as expected output; inputting the image to be processed into the trained model, and outputting a high signal-to-noise ratio image. The application can effectively separate disease signals in a strong noise interference background under single disease scene and combined disease scene conditions, and does not appear signal distortion, amplitude attenuation and edge blur, and the denoised image is almost consistent with the original clean image, thereby providing a strong basis for subsequent intelligent disease recognition.

The present application provides a kind of coastal water level prediction method based on multi-modal observation data, belong to coastal water level prediction technical field, the present application is by collecting the multi-modal marine observation data of different sensors, the spatial and temporal alignment of heterogeneous data is realized by establishing multi-scale space-time registration matrix, constructs marine dynamic process feature extractor to identify astronomical tide storm surge and wave characteristics and calculate nonlinear coupling parameters, using adaptive space-time fusion algorithm according to the weight of data quality dynamic adjustment generation space-time consistency dataset, establishes marine dynamic coupling strength discrimination model to determine modeling strategy, finally generates the coastal water level prediction product containing prediction value confidence interval and risk warning, solves the technical problem that multi-modal marine observation data is difficult to effectively fuse under the condition of space-time scale mismatch, resulting in insufficient coastal water level prediction accuracy.

The invention provides a high-efficiency hyperspectral remote sensing image generation method based on spatial spectrum information self-selection. The method mainly comprises the following three core steps: 1, multi-level self-selection spatial feature coding; 2, multi-scale semantic fusion based on mixed pooling; and 3, spatial prior guidance and waveband relevance modeling. Starting from the characteristics of a hyperspectral remote sensing image generation task, a generation framework guided from spatial characteristics to spectral information driving is constructed, and a spatial-spectral information self-selection mechanism and a lightweight calculation strategy are introduced, so that the model can adaptively and dynamically pay attention to important spatial regions and spectral bands. The problems of insufficient physical consistency, redundant feature information, high video memory occupation and the like of a current hyperspectral remote sensing image generation method are effectively relieved, and the method has the advantages of high reconstruction precision, low calculation complexity, high training reasoning speed and the like; and high-quality hyperspectral remote sensing image data support can be provided for remote sensing downstream tasks such as image classification and change detection.

The application discloses a coal mine image segmentation model and method based on a VMamba and multi-expert hybrid network and a construction method thereof. The coal mine image segmentation model is constructed. The image block encoding layer output of an encoder is taken as the input of the first encoding layer of a first VSS network, and the output of the first encoding layer of the first VSS network is taken as the input of the output end feature learning module. The input of the output end feature learning module is taken as the input of the first decoding layer of a decoder, and the output of the first decoding layer of the decoder is taken as the input of the second decoding layer of the decoder. The input of the second decoding layer of the decoder is taken as the input of a segmentation head, and the output of the segmentation head is taken as a segmented image. The application can significantly improve the segmentation precision and calculation efficiency of the coal image.

The application discloses a method for establishing a brain glioma grading hybrid network based on multi-view feature fusion, which comprises four steps of data preprocessing, feature extraction, multi-view feature fusion and grading processing of the obtained features through a classifier; the method can balance global receptive field and induction bias, combines global and local multi-scale and multi-granularity features, balances global receptive field and induction bias at the same time, and enhances feature extraction capability; by constructing different view data sources (interlayer view, scanning plane view and modality view) and designing a novel feature fusion module, the network can fully and effectively utilize key information provided by different scanning planes and different modalities of MRI, including interlayer, spatial and multi-modality information, and realizes comprehensive cognition of brain glioma.

The application discloses a remote sensing image target detection method and system based on an optimized SSD algorithm, first, inputting a to-be-detected remote sensing image into an efficient remote sensing image target feature extraction network to extract key features, and obtaining a required feature map; then inputting the obtained feature map into a forward and reverse iterative fusion multi-scale feature network to perform feature fusion, and generating a new feature map; finally, utilizing an anchor box matching network based on a clustering algorithm to perform anchor box clustering matching on different categories of targets in the image according to the new feature generated by fusion, so that each to-be-detected target obtains the most suitable anchor box as a target detection result; the original feature extraction network VGG16 in the SSD algorithm is optimized, an improved RFB module and an efficient attention mechanism are added into the network, the feature extraction capability of the network is improved, the anchor box matching network based on the clustering algorithm is integrated into the algorithm, the accuracy of the anchor box matching target of the algorithm is improved, and therefore the detection capability is improved.

This invention discloses a rapid detection method and system for multiple types of hidden dangers in concrete dams based on the HDIOCNet model. First, infrared-visible dual-light images of the concrete structure surface are acquired, preprocessed, and semantically annotated to construct a training dataset integrating the dual-light images for concrete dam hidden danger detection. Second, a deep learning model, HDIOCNet, for rapid detection of multiple types of hidden dangers in concrete dams is constructed. This model employs an efficient encoder-decoder hourglass structure. The encoder part includes a lightweight edge and texture feature extraction backbone network and a multi-scale receptive field extended dilated convolutional pooling pyramid to encode the features of the input dual-light images and output feature maps at different semantic levels. Finally, the HDIOCNet model is trained based on the dual-light dataset. The trained model is used to predict the types of hidden dangers and calculate and compare the areas of hidden dangers. This invention has the advantages of high detection accuracy, fast detection speed, and intelligence, providing an efficient solution for the safety monitoring of concrete structures.

This invention belongs to the field of image classification technology and relates to a method and system for tempered glass inspection based on a bi-branch inverse residual network model. The bi-branch inverse residual network model includes a preprocessing unit, a first feature extraction unit, a second feature extraction unit, a third feature extraction unit, a fourth feature extraction unit, and a classification unit connected sequentially. The feature extraction unit is composed of one or more bi-branch inverse residual modules. The model provided by this invention has advantages such as high accuracy, few parameters, and low computational cost. When used for tempered glass quality inspection, it can improve inspection efficiency and accuracy, while also ensuring non-destructive and continuous inspection, thus guaranteeing the safety and efficiency of glass production.

The application relates to the technical field of target detection, in particular to a high-altitude electric power operation violation identification method, which discloses the following steps: collecting high-definition images and low-quality images of an electric power operation site by using a drone, and making an electric power site high-definition data set; constructing an image enhancement network based on a local information enhancement module and a global information enhancement module; building a violation identification model based on an improved YOLOv8; selecting a suitable optimizer and a loss function; training and optimizing the constructed model. The optimal model after training and optimization is converted into an RKNN model in the front end, and is deployed to the front end of the drone. Real-time detection is realized by relying on the camera in the front end, and whether the operation personnel violate the rules is judged by the judgment unit according to the detection result. The application provides a high-altitude electric power operation violation identification method, which considers the construction of an operation data set, the enhancement of low-quality images, and simultaneously proposes a violation identification network model based on an improved YOLOv8, so that the accuracy of detection is maximally improved.