459 results about "Feature evaluation" patented technology

The invention discloses an industrial odor online monitoring system based on big data. The system comprises a data acquisition module, a preprocessing module, a feature fusion module, a multi-source sensing weight analysis module, a dynamic risk assessment module, a self-adaptive correction module and a real-time feedback module. Data interaction among the modules adopts a block chain encryption transmission protocol, and a feature evaluation mapping chain with a timestamp is established in a distributed database. Through the innovative design of heterogeneous sensor network collaborative perception, cross-modal feature organic fusion, pollution diffusion three-dimensional dynamic deduction and a credible data governance architecture, a new industrial bad smell monitoring normal form with environment intelligent adaptability is constructed. The four technical effects form complementary support from four dimensions of data acquisition credibility, feature modeling scientificity, risk assessment accuracy and system decision robustness.

The invention relates to a large-scale data mining method based on multi-tuple data optimization. The method comprises the following steps: firstly, carrying out importance analysis and sorting on each dimension feature in multi-tuple data by designing a feature evaluation mechanism based on information gain and mutual information; on a data set after feature selection and extraction, an improved clustering algorithm is applied, the advantages of density clustering and hierarchical clustering are combined, and a cluster of a structure is recognized in high-dimensional multi-tuple data; based on a cluster obtained through clustering analysis, a cross-dimension association rule mining algorithm is designed, and a cross-dimension association relationship is revealed in the multi-tuple data; constructing a hybrid prediction model fusing multiple regression and deep learning technologies; and the prediction accuracy of the target variable is improved by utilizing the multi-dimensional information in the multi-element group data and the mined association rule.

The present invention discloses a health status assessment method and system for multi-source heterogeneous power data. The health status assessment method and system for multi-source heterogeneous power data include: S110, constructing a unified data model for multi-source heterogeneous power data; S120, extracting multidimensional data features based on the unified data model; S130, constructing a health status assessment model based on the extracted data features; S140, establishing a real-time health status monitoring and alarm mechanism based on the health status assessment model; the present application provides a real-time monitoring and alarm mechanism for the health status of multi-source heterogeneous power data, accurately detecting abnormal situations of power data through a multi-feature assessment model, and providing convenience for operation and maintenance personnel.

The invention provides a telecommunication service fraud-related risk security assessment system based on a multi-modal fusion model, and the system comprises a multi-source data collection assembly which is responsible for obtaining original telecommunication service data in multiple ways, carrying out the preprocessing of the original telecommunication service data, and obtaining first telecommunication service data; performing feature extraction and behavior pattern analysis on the first telecommunication service data to obtain key features related to the telecommunication service; the assessment model construction component is responsible for constructing a fraud-related risk assessment model based on machine learning, inputting the key features into the fraud-related risk assessment model, outputting an intelligent assessment control matrix, classifying and rating fraud-related risks, and generating a risk assessment result; and the service collaborative linkage assembly is responsible for early warning the risk level of the telecommunication service system according to the risk assessment result, and taking prevention measures according to the risk level. According to the method, an objective evaluation standard is established, and the conversion of risk identification from experience judgment to data driving is realized.

The invention discloses a water and soil loss telemetry data analysis method and system based on deep learning, and relates to the technical field of data analysis, and the method comprises the steps: extracting soil loss data, surface runoff data and topographic data; fusing the soil loss data, the surface runoff data and the topographic data, and obtaining water and soil loss risk index data and soil erosion rate data through a feature extraction technology; calculating weight values of the water and soil loss risk index data and the soil erosion rate data by using a feature evaluation technology; and based on the water and soil loss risk index data, the soil erosion rate data and the weight value, constructing a water and soil loss prediction model by using a deep learning framework and a space-time analysis algorithm. According to the invention, through telemetry data collection and feature extraction, key water and soil loss information can be rapidly obtained, the data processing flow is simplified, and the accuracy and reliability of data are ensured.

The invention relates to the technical field of robots, and discloses a robot nondestructive testing method and system based on artificial intelligence, and the method comprises the steps: obtaining multi-modal sensing data and motion state parameters; constructing a three-dimensional material feature model and analyzing to generate an initial defect feature evaluation model; loading material attributes and detection precision constraints to generate a comprehensive model, and combining motion parameter simulation to obtain path regulation and control data; training the dynamic defect prediction model to obtain a detection path correction model and generating initial detection parameters; obtaining dynamic correction simulation information based on the initial parameters and the like; constructing a multi-objective optimization model, and optimizing to obtain an optimal detection path parameter; and generating a safety detection probability by combining real-time model reasoning, and adjusting a detection strategy to realize dynamic path matching. The system comprises a data acquisition module, a feature modeling and analysis module, a comprehensive evaluation and simulation module and the like. The method improves the detection precision, efficiency and adaptability, and is suitable for nondestructive detection of complex targets.

The invention relates to the technical field of data processing, and discloses a seawall feature information extraction method and system based on multi-source remote sensing data. The method comprises the following steps: carrying out wave band combination enhancement, geometric correction and radiation correction on a multi-source remote sensing image to obtain standardized data; extracting a seawall target by using multi-scale segmentation; recognizing structural features through profile analysis and spectrum matching; calculating vegetation indexes and texture features to obtain ecological information; establishing a precision evaluation system; and finally constructing a feature information database. Through the technical means of multi-source data fusion processing, multi-scale feature extraction, ecological feature evaluation and the like, the precision and efficiency of seawall feature information extraction are improved, and the problems existing in the aspects of data acquisition, processing and evaluation in a traditional method are solved.

The invention relates to a radiographic image quality evaluation and enhancement method and system based on deep learning. The method comprises the following steps: constructing a multi-modal radiographic image quality evaluation data set containing noise types, artifact levels and diagnostic availability labels; designing a double-branch deep neural network architecture containing a quality evaluation branch and an image enhancement branch, the quality evaluation branch fusing global and local features to evaluate image quality, and the image enhancement branch realizing local enhancement of a focus area; coordinating a loss function optimization process by adopting a self-adaptive joint training strategy; and deploying an online correction module to iteratively optimize the enhancement parameters. The system comprises an image acquisition interface module, a preprocessing module, a quality evaluation engine module, an intelligent enhancement module, a post-processing module and the like. The technology can accurately evaluate the quality of the radiation image, effectively enhance the image, reduce noise and artifact interference and improve the diagnosis availability, and has important application value in the field of medical image processing.

The invention provides an image noise mark feature selection method and system, a storage medium and a computer. The method comprises the steps of obtaining a to-be-processed image noise mark data set; embedding a sample set in the image noise mark data set into a multi-granularity fuzzy cluster to construct a dynamic fuzzy membership evaluation matrix; dynamically evolving a multi-level high-precision granular ball cluster; obtaining mark distribution with high identification degree; constructing a rough perception feature evaluation framework based on granular ball topology driving, extracting decision equivalence classes by combining rough set upper and lower approximation and extended positive domain models, and determining and measuring the contribution degree of each feature to a decision system by fusing multi-granular-ball decision boundary information based on a dependency degree quantitative model; a particle and ball structure consistency verification mechanism is introduced, and multi-level evaluation is carried out on the importance of the features through dependency and consistency. According to the method, the optimal feature subset with strong anti-noise performance and high discrimination capability is obtained, and stable and efficient input support is provided for a subsequent image noise mark learning model.

The invention provides a sound standardization evaluation system for multi-dimensional acoustic parameter and emotional scene dynamic analysis, and aims to realize structured and objective evaluation of voice performance. The system comprises an audio acquisition module, an audio processing module, a standard evaluation module and an integration processing module. The audio processing module performs noise reduction, pre-emphasis, sampling rate adjustment, framing and slicing operation on the original voice data; the standard evaluation module executes character and pronunciation accuracy analysis, acoustic basic skill feature evaluation or emotion and scene adaptation degree analysis based on multi-modal fusion according to an evaluation type set by a user; parameters including fundamental frequency, formant, sound intensity, fundamental frequency perturbation, amplitude perturbation and the like are extracted in evaluation, and classification prediction is completed in combination with a machine learning model; and finally, the integration processing module generates an evaluation report including a scoring result, question prompts and personalized training suggestions, and provides closed-loop feedback to support sound training and expression optimization.

The invention discloses an indoor air quality monitoring method and system based on multi-dimensional analysis, and belongs to the technical field of air quality monitoring. The indoor air quality monitoring method based on multi-dimensional analysis comprises the steps that various data of a chemical sensor, an optical sensor and an acoustic sensor are collected; preprocessing the data, including unifying sampling frequency and normalizing different data formats; performing multi-dimensional feature extraction to obtain a multi-modal feature; carrying out feature fusion by adopting a machine learning algorithm; the air quality state is evaluated based on the fusion features; through multi-source data fusion analysis, the problems of one-sided monitoring result and insufficient reliability of a single sensor are solved, the monitoring accuracy and reliability are improved, and the method can be widely applied to scenes such as smart homes and office buildings.

The invention provides a packing material defect detection method and system based on machine vision, and relates to the technical field of machine vision, and the method comprises the steps: collecting a multi-angle image sequence to construct a depth map, and carrying out the geometric correction and illumination compensation of an image; generating a feature map sequence by using a multi-scale direction filter bank and a self-adaptive direction enhancement factor, and constructing a feature pyramid structure to realize hierarchical segmentation of a flaw region; extracting a multi-dimensional feature set containing local textures, gray statistics and shape description, distributing weight coefficients through a feature evaluation matrix, and fusing to generate a feature vector; a double-order cascade classifier is used for classification, a lightweight decision tree is used for rapid pre-classification in the first order, and features are extracted and fused through an attention-enhanced deep network in the second order. According to the invention, the accuracy and real-time performance of the defect detection of the packing material are improved, and the adaptability of the detection system to different types of defects is enhanced.

The invention relates to the technical field of information security, in particular to a sensitive information privacy calculation multi-dimensional security risk evaluation method, which comprises the following steps of: cooperatively collecting sensitive data information by setting a plurality of distributed nodes, and constructing the collected sensitive data information into corresponding feature vectors; performing privacy calculation based on the constructed feature vector; performing multi-dimensional assessment of the security risk based on a recovery result of the feature vector; according to the method, the feature vectors are encrypted layer by layer by using a multi-stage differential privacy mechanism, original data exposure is avoided, the risk of privacy disclosure is reduced, and personalized encryption is realized by dynamically adjusting the noise variance, and privacy protection and data validity are both considered; meanwhile, a distributed dynamic key management and decryption mechanism is introduced, and the key is prevented from being maliciously inferred through multi-round decryption calculation; and dynamically dividing risk level intervals, automatically matching feature evaluation results, and quickly identifying risk levels.

The invention discloses an electromagnetic bearing fault mode analysis method for fault diagnosis of a primary helium fan. The method comprises the following steps: establishing an electromagnetic bearing-rotor-rack high-fidelity coupling dynamic model; extracting fault feature frequency and vibration amplitude, preprocessing the monitoring signal, and extracting multi-source heterogeneous fault features; evaluating the sensitivity of each feature to a fault, designing an adaptive weighted D-S evidence theory fusion algorithm, and identifying a bearing fault mode by dynamically adjusting the weight of an information source; building a convolutional neural network model to automatically extract deep fault features corresponding to the working conditions, and performing diagnosis model adaptation among different working conditions by adopting a transfer learning strategy; the GRU network is used for training and learning time sequence monitoring data, the development trend of bearing faults is predicted, and a heuristic algorithm is used for optimizing a fan operation and maintenance strategy. According to the method, the limitation of the existing method can be overcome, accurate diagnosis and trend prediction of the bearing fault are realized, and key technical support is provided for intelligent operation and maintenance of the rotating machinery.

The invention discloses a heart failure patient remote monitoring and early warning system based on the Internet of Things, and the system comprises a data collection and preprocessing module which is used for collecting and preprocessing multi-source physiological and behavior data, and generating an input sample set; the feature evaluation module is used for constructing an extreme gradient boosting tree algorithm model and obtaining feature importance scores; the feature screening module is used for executing a recursive feature elimination process and outputting an optimal feature subset; the risk judgment module is used for inputting the optimal feature subset into an extreme gradient boosting tree algorithm model to generate a heart failure risk score; the early warning triggering module is used for triggering early warning and sending early warning information to the patient and the platform; the feedback recording module is used for generating feedback information; and the dynamic optimization module is used for updating the input sample set based on the feedback information, retraining the extreme gradient boosting tree algorithm model and screening the features. According to the method, the extreme gradient boosting tree and recursive feature elimination are fused, and accurate evaluation and dynamic early warning of the multi-source data of the heart failure patient are realized.

The invention discloses an accurate feeding system and method based on machine vision. The system comprises an image acquisition module, an image processing module, a satiation degree evaluation module, an automatic feeding machine and a control module. The image acquisition module is used for acquiring a water surface spray image of the fishpond in real time; the image processing module is in signal connection with the image acquisition module and is used for carrying out segmentation and feature extraction on the water surface spray image acquired in real time by adopting a convolutional neural network to obtain spray features of the water surface spray image, and the spray features comprise the size and the duration of the spray; the water spray features before feeding, during feeding and after feeding are stored in a feeding water spray image database; and the satiation degree evaluation module is used for evaluating the ingestion active degree and the satiation degree of the fish school according to the ingestion spray characteristics on the basis of a satiation degree evaluation model. According to the method, the feeding speed is adjusted according to the collected and analyzed water spray characteristics on the water surface, so that the feeding speed is coordinated with the feeding speed, precise feeding is achieved, and residual feed is reduced.

The invention discloses an underwater robot real-time three-dimensional map reconstruction system and method fused with an SLAM algorithm, and belongs to the technical field of three-dimensional map reconstruction. By constructing a confidence coefficient system of three-dimensional points, a spatial trajectory fluctuation index and an inter-frame texture change rate are introduced as dynamic feature evaluation indexes; and a regression model is adopted to score and grade the severity of the false points, and differential processing of different grades of point clouds is realized in combination with a dynamic suppression mechanism, so that an unstable structure is effectively eliminated, and the stability and reliability of the sparse three-dimensional map are remarkably improved. Therefore, the positioning precision, the path planning capability and the task execution safety of the underwater robot in a complex environment are enhanced.

The invention provides a dynamic data traceability method and system oriented to network security, and relates to the technical field of network security and data governance crossing, and the method comprises the steps: carrying out the real-time visual monitoring of a user interaction interface of a data transaction platform, and capturing the visual behavior information related to data calling; extracting feature parameters based on the visual behavior information to obtain a visual feature data set, and performing structured conversion on the visual feature data set to obtain a structured call log record containing visual behavior associated information; based on the structured call log record, extracting a time sequence, a frequency and a spatial distribution feature of a call behavior as basic indexes; a plurality of feature sampling states are dynamically selected in a monitoring time window, and a dynamic feature evaluation set is constructed based on the feature sampling states. According to the invention, the accuracy and controllability of network security protection in a data transaction scene can be improved.

The invention discloses a tracking and tracing method and system based on an RFID Internet of Things technology. The method comprises the following steps: firstly, realizing full-link abnormal risk quantification through multi-source data fusion and dynamic evaluation; the method comprises the following steps: carrying out feature embedding processing on RFID full-link sensing data of a tracked article, and extracting a sensing feature vector; generating fusion depth features reflecting full-link association through depth feature fusion; evaluating the fusion depth feature based on a multi-traceability link to obtain the abnormal confidence of each link; analyzing the perception feature vector to obtain an influence coefficient of each link on the whole link abnormity; and dynamically aggregating the abnormal confidence coefficient and the influence coefficient, and calculating a full-link abnormal probability. According to the method, the problems of multi-source data time sequence alignment and feature fusion are solved, accurate positioning of abnormities of all links and dynamic evaluation of full-link risks are achieved, the method is suitable for full-link tracking scenes such as agricultural products and aquatic products, and the quality control efficiency is improved.

The invention discloses an intelligent e-commerce user abnormal behavior detection method and system. The method comprises the steps of data acquisition, data preprocessing, data processing and e-commerce user abnormal behavior detection. The invention belongs to the field of anomaly detection, and particularly relates to an intelligent e-commerce user abnormal behavior detection method and system, and the method comprises the steps: filtering a part of abnormal behaviors in advance through designing low-fluctuation users, sudden abnormal users and diverse abnormal users; based on design distance measurement, the similarity of user behavior sequences is measured, and secondary optimization is carried out on the centroid, so that the adaptability of the centroid to user behavior distribution is improved; the accuracy of final user abnormal behavior detection is improved; by designing a dynamic distance threshold and multi-dimensional feature evaluation, comprehensively considering the user behavior data, and dynamically adjusting an anomaly detection threshold according to the distribution characteristics of the user behavior data, an abnormal user is accurately identified, the misjudgment risk is reduced, and the detection accuracy is improved.

The invention provides a monitoring data linkage method between code-free applications, automatic identification and mapping of data structures between the applications are realized by constructing a metadata description model and a field similarity function, and data change detection is performed by adopting an improved locality sensitive hash algorithm. Through a distributed transaction control mechanism, data consistency in cross-application data linkage is ensured, and a multi-stage bloom filter and a distributed transaction feature evaluation model are designed. Furthermore, data synchronization is realized by utilizing a synchronous vector field, and data synchronization batch processing and concurrency adjustment are optimized in combination with a Gaussian divergence theorem. Besides, the invention provides abnormal recovery path optimization based on the minimum action quantity, the reliability of data linkage is ensured, and a dynamic data security protection mechanism is constructed through a wave function, a diffusion equation and nonlinear mapping.

The invention relates to the technical field of image processing, in particular to an AI-based graphic design optimization system, which comprises an information entropy distribution evaluation module, an information entropy distribution evaluation module, an information entropy evaluation module, an information entropy evaluation module, a feature evaluation module, a feature evaluation module, a feature evaluation module and a feature evaluation module, wherein the information entropy distribution evaluation module acquires information of each element in a design graph, calculates information entropy of a feature item of each element to acquire an information entropy value of the feature and judges visual importance difference among the elements; and sorting the elements according to information importance, and generating an importance sequence of the design elements. According to the method, the feature information of each element in the design graph is analyzed, the visual importance of the elements is measured by the information entropy, and the size, the position, the color and the like of the element information are quantified and sorted, so that the design element layout is ensured to have a clear visual level. Therefore, the priority of each design element is clearer, and the user can pay attention to the important elements first in the design work. Through harmony of color features such as hue, brightness, saturation and the like, optimized color matching is calculated, and the visual harmony of the design is improved.

The invention discloses an education evaluation assisting method and system based on artificial intelligence. The method comprises the steps of data collection, data optimization, learning behavior evaluation, education feature evaluation and education evaluation assisting. The invention relates to the technical field of education management, in particular to an education evaluation auxiliary method and system based on artificial intelligence. According to the scheme, a two-way correlation analysis method combining learning behaviors and education features is adopted, preliminary analysis is conducted through learning behavior evaluation, multi-dimensional extraction is conducted through the education features, and the education evaluation accuracy is improved. The analysis depth and the analysis specificity of education evaluation are improved, and meanwhile, an effective way for assisting the education evaluation by an intelligent method is provided; performing learning behavior evaluation by adopting a method of combining a causal enlightenment enhanced graph nerve long-short-term memory integration model; and carrying out educational feature evaluation by adopting a graph neural network method combining knowledge migration and a knowledge graph.

The invention discloses an LSTM (Long Short Term Memory)-based fraud-related network traffic identification method and a server, and is applied to the field of network traffic analysis. According to the method, original flow data are collected in real time through network flow collection equipment, communication feature data and behavior feature data are extracted from each communication address, and weighted combination is carried out by adopting adaptive weight coefficients based on feature matching degree and feature abnormality to obtain a suspicious degree score; and outputting an early warning result to the communication address exceeding the preset suspicious threshold. According to the technical scheme provided by the invention, a multi-layer fraud-related network flow identification architecture and a closed-loop feature evaluation mechanism are constructed, so that the identification accuracy of continuously changing fraud-related communication behaviors is effectively improved.

The invention discloses a Parkinson's gait assessment method and system based on millimeter wave radar signals. The method comprises the following steps: acquiring a millimeter wave radio frequency signal in an environment, and performing data processing on the signal to obtain a distance spectrum of the change of the position of the body of the Parkinson patient and a Doppler spectrum reflecting the change of the physical degree of the Parkinson patient; extracting a peak value of the distance spectrum at each moment through a filtering method, calculating a cross correlation coefficient of a Doppler spectrum as an index of gait mode matching, and segmenting the Doppler spectrum according to each step; gait feature calculation is carried out according to the filtered distance spectrum peak value and the segmented Doppler spectrum; and normalizing the gait features obtained by calculation, and inputting the normalized gait features into the trained machine learning model for feature evaluation to obtain the UPDRS-III score of the patient. According to the method, non-invasive measurement and evaluation of the Parkinson's gait characteristics are achieved in a low-cost mode, operation is convenient, the evaluation effect is accurate, and quantitative data and diagnosis reference are provided for doctors.

The invention discloses a virtual power plant load prediction system based on big data, and relates to the technical field of power load prediction. The virtual power plant load prediction system based on big data comprises a time sequence evaluation module, a numerical characteristic evaluation module, a data characteristic evaluation module and a multi-source data evaluation module. According to the method, the time sequence abnormal index is obtained through the obtained time sequence data, whether a time adjustment measure is taken or not is judged, and then the numerical characteristic abnormal index is obtained through the obtained numerical characteristic data to judge whether a data adjustment measure is taken or not; then, a data feature deviation index is obtained through the obtained data feature data to judge whether to regulate and control the data to adjust the measure strength, and finally, a multi-source data consistency evaluation index is obtained through the obtained multi-source feature data to judge whether to carry out equipment operation and maintenance, so that the effect of obtaining more accurate data to carry out power load prediction is achieved; the problem that a data source is unstable during power load prediction in the prior art is solved.

The invention relates to the technical field of data processing, and discloses an intelligent talent management method and system for chain enterprises, and the method comprises the steps: carrying out the multi-dimensional data collection and preprocessing of an internal system and an external data source of an enterprise, and obtaining a multi-dimensional employee data set; performing talent network construction, and fusing personal attributes and network attributes to obtain a multi-level talent distribution network; extracting multi-dimensional evaluation indexes according to the multi-dimensional employee data set and carrying out weight distribution to obtain an evaluation index system; performing feature index calculation on the multi-dimensional employee data set according to the evaluation index system to obtain a feature evaluation index set; and performing graph convolutional neural network and multi-scale aggregation analysis on the feature evaluation index set and the multi-level talent distribution network to obtain a plurality of talent recognition results, and matching an intelligent management strategy of each talent recognition result, so that intelligent talent management of chain enterprises is realized.

The invention discloses an automobile circuit board drilling burr detection method and system based on area growth, and aims to improve the quality control efficiency and accuracy of automobile circuit board manufacturing. The system is used for quickly and accurately identifying and evaluating burr defects in the automobile circuit board drilling process; a high-definition image and three-dimensional scanning data of an automobile circuit board drilling area are obtained through a high-precision image acquisition device, and image preprocessing including noise reduction, enhancement and graying processing is carried out. An intelligent seed point selection strategy is adopted, three-dimensional data are combined, deep analysis is carried out on the areas, and various features including the shape, the size, the brightness and the like are extracted. According to the method, the extracted features are classified by training a primitive machine learning model, the burrs are effectively identified and classified, the positions of the burrs are accurately positioned, the potential influence on the performance of the automobile circuit board is evaluated according to the features, a detailed detection report is generated, and a scientific basis is provided for quality control and production optimization.

The invention provides an efficient data retrieval method and system oriented to a large language model. The method comprises the following steps: determining an intention vector set matched with a user intention; determining a specified number of candidate text fragments with the highest correlation with the user intention in an index database; obtaining multi-dimensional feature evaluation data of each candidate text fragment, and constructing a multi-dimensional feature numerical matrix of each candidate text fragment according to the multi-dimensional feature evaluation data; according to the multi-dimensional feature numerical value matrix, selecting selected text fragments from the candidate text fragments; sorting the selected text fragments, and sequentially performing scene matching on the selected text fragments and a pre-cached high-frequency query scene according to the sorting, so as to extract cache content matched with the target high-frequency query scene from the cache as a query result when a matching result represents that the current query scene accords with the target high-frequency query scene. The invention provides an efficient data retrieval method, and the quality and precision of data retrieval are improved.

The invention relates to the technical field of underwater detection, and discloses a forward-looking sonar underwater detection system based on multi-beam forming. The system comprises seven modules: a sonar data acquisition module obtains a multi-beam sonar original data sequence; a historical sonar database retrieval module retrieves a matched historical detection record set according to the region identifier; the heterogeneous sonar record distinguishing module classifies the historical sets and determines a plurality of distinguishing sets; a sonar feature extraction module traverses and distinguishes the set calculation to obtain a sonar feature set value and an associated bandwidth; the multi-scale sonar analysis module processes a network layer receptive field by taking a feature association bandwidth as a beam, and analyzes original data to obtain a multi-scale detection feature set; a cross-scale sonar fusion module performs interactive fusion on the multi-scale features to generate a target interactive fusion sonar feature set; and the underwater target area evaluation module outputs a thermodynamic map containing potential target area probability distribution based on fusion feature evaluation, and adapts to complex underwater environment detection.