29results about How to "Improve characterization accuracy" patented technology

The invention discloses a multi-parameter distortion compensation method for a fractured rock mass seepage similar model test, and relates to the technical field of rock mass engineering physical model tests. The method comprises the following steps of: firstly, determining two constraint conditions that the fluid property is invariable and the gravitational acceleration is invariable, correspondingly selecting a basic physical quantity and extracting a core criterion; then analyzing three types of distortion sources of geometric dimensions, dielectric materials and boundary conditions, and establishing a distortion function according to result parameters of seepage flow velocity, seepage flow and seepage pressure; secondly, parameters difficult to control are selected as distortion items, parameters easy to regulate and control are selected as compensation items, the distortion items are substituted into a distortion function to obtain a distortion ratio, the optimal distortion ratio of the compensation items is solved based on a compensation balance formula, and the parameters are adjusted; and finally verifying the interval range of the actually measured distortion ratio. According to the method, the multi-source distortion coupling influence is accurately quantified, the compensation pertinence and precision are improved, operation is easy and convenient, universality is high, and reliable support is provided for underground engineering seepage research.

The application provides a hypersonic platform radar target detection method, device and electronic equipment, relates to the technical field of hypersonic radar target detection, and the method comprises the following steps: inputting a received first echo signal into a time domain echo signal model, converting the first echo signal from a first signal domain to a second signal domain by using the time domain echo signal model, and obtaining a second echo signal; filtering the second echo signal by using an optimal parameter filter, obtaining a filtered third echo signal, converting the third echo signal into a third signal domain, and obtaining a coherent accumulation result; determining whether a target is detected by using a constant false alarm rate detection based on the coherent accumulation result, and determining the distance of the target based on the position of a peak value in the accumulation result in the case that it is determined that the target is detected. The hypersonic platform radar target detection method, device and electronic equipment provided by the application can realize high-precision, high-efficiency and highly adaptable radar target detection.

The present application relates to the technical field of big data analysis, in particular to a regional power distribution network short-term load intelligent prediction system based on big data, which comprises a frequency domain filtering and load mode extraction module, which extracts a high-frequency fluctuation load sequence and a low-frequency reference load sequence according to the power distribution transformer topology node position coordinates and the initial power sequence collected by the smart meter at high frequency. In the present application, the low-frequency reference load sequence is trend extrapolated, and then echoed with the aforementioned disturbance result, so that the long-term electricity consumption trend and the short-term external impact jointly enter the prediction process, avoiding weakening the instantaneous fluctuation by only relying on the overall historical trend, and avoiding ignoring the basic load inertia by only focusing on the instantaneous change. Thus, synergistic gains are formed in terms of layered expression of load changes, depiction of node linkage relationship, depth of meteorological disturbance mapping and integrity of future trend connection, which can improve the adaptability of the regional power distribution network short-term load prediction result to complex scenarios.

The invention provides a municipal road maintenance demand prediction method and system based on machine learning, and relates to the technical field of machine learning, and the method comprises the steps: obtaining the service starting time of a municipal road and the passing frequency and full load proportion of heavy-load vehicles in a plurality of sampling periods; difference calculation is carried out on the service starting time and the sampling time points in each sampling period, and a space-time incidence matrix is constructed; taking an attenuation coefficient in the space-time incidence matrix as a dynamic regulation factor to obtain a bit stream sequence; mapping the bit stream sequence as an observation sequence to a preset road grid map, and determining an optimal transfer path through optimal path planning; and determining a grid step length for transferring the road state to the maintenance state at the current moment from the optimal transfer path, and determining maintenance demand prediction information comprising a maintenance priority and a predicted maintenance date. The precision, objectivity and foresight of municipal road maintenance demand prediction are realized.

This invention discloses a planning method and system for a hydrogen-containing integrated energy system. The method includes: constructing equipment instance models of various energy types in the system by configuring parameters; wherein the equipment includes electrolyzer equipment; constructing a load model in the system by configuring the load demand of the area to be planned, the load demand supporting configuration at multiple time scales; obtaining the energy price model of the area to be planned and completing the energy price configuration; constructing a system topology by graphically associating the equipment instance models, load models, and energy price models through a visual configuration interface, and configuring constraint rules on the system topology; wherein the constraint rules include start-up and shutdown optimization constraints for the electrolyzer; transforming the system topology with configured constraint rules into a mixed-integer linear programming model and performing optimization to obtain the configuration scheme and operation strategy of each device in the system. This invention solves the problems of insufficient characterization of equipment operating losses and low modeling efficiency in existing planning methods by introducing refined electrolyzer start-up and shutdown constraints and a visual modeling process, thus achieving efficient and accurate planning for hydrogen-containing integrated energy systems.

The invention discloses a blockage detection method for a particulate filter, and relates to the field of vehicle tail gas treatment, and the method comprises the following steps: generating a pressure difference volume flow relation curve by using a preset curve fitting model constructed based on a weighted least square method, calculating the change trend similarity between the pressure difference volume flow relation curve and a curve in a non-blockage state, and determining the blockage of the particulate filter according to the change trend similarity. And based on a comparison result corresponding to the change trend similarity, sending a cloud detection confidence coefficient to the vehicle end controller, so that the vehicle end controller superposes the cloud detection confidence coefficient and the vehicle end detection confidence coefficient, and outputting a jam detection result. According to the method, the disturbance variable introduced by sensor drift in the pressure difference and the volume flow is suppressed based on the weighted least square method, the precision of the generated pressure difference volume flow relation curve is improved, the blockage detection is realized based on the change trend similarity, the cloud detection confidence coefficient and the vehicle end detection confidence coefficient are further superposed by the vehicle end controller, and the vehicle end detection accuracy is improved. Double verification is realized, and the blockage detection precision of the particle trap is improved.

The invention belongs to the field of network social group analysis, and provides a multi-expert collaborative network social group description method based on modal dynamic fusion. The method comprises the following steps: firstly, inputting multi-modal original data of a network social group, and respectively capturing multi-modal private feature representations through a modal representation adaptive extraction module; then, deep fusion of cross-modal information is realized by utilizing a bidirectional state space model through a modal complementary information fusion module, and unified group multi-modal fusion representation is generated; thirdly, a KAN multi-expert network architecture is introduced through a multi-expert collaborative prediction module, different experts are adapted to diversified group characteristic modes such as mainstream, small crowds and temporary groups, and dynamic weights are calculated through a gating network to achieve accurate fusion of expert output; and finally, in combination with long-tail boundary perception loss function optimization model training, outputting a multi-dimensional feature description result of the group. According to the method, more comprehensive and accurate group feature representation can be obtained, and the depicting performance of the network social group in a complex scene is improved.

The application discloses a geological comprehensive information analysis system for exploration, comprising: a data acquisition module: acquiring general characteristic data of an exploration area collected by a multi-source sensor in real time; a probability density estimation unit: obtaining a probability density function of the general characteristic under a corresponding geological type for each predefined geological type; a continuous type representation unit: inputting the general characteristic data acquired by the data acquisition module into a kernel density estimation model of all geological types, generating a fuzzy membership vector for continuously describing the geological attribution of the current exploration area, and taking the fuzzy membership vector as a continuous type representation; an anomaly threshold database: storing judging standards according to different geological types; and an anomaly judgment module: dynamically determining an anomaly threshold matched with a current exploration geological state from the anomaly threshold database based on the continuous type representation, comparing real-time fusion data with the anomaly threshold, judging whether an anomaly exists in the current exploration geology, and triggering an early warning.

The application discloses a chloride-based multi-parameter humidity and conductivity comprehensive calculation method, comprising the following steps: S1, collecting basic parameter information of chloride materials, and performing relevant data arrangement, and comparing after entering a system; S2, then setting an environmental humidity change range, and obtaining moisture absorption and surface water layer thickness under different relative humidity conditions; S3, establishing a relevant coupling model, and considering ion mobility, charge transmission channel number and interface polarization effect of the chloride materials under different humidity; S4, introducing a multi-parameter regulation factor, correcting environmental temperature, pressure and material surface state, and obtaining correct data; the method has high application prospect and popularization value.

The embodiment of the invention provides a transverse wave static correction value determination method and device, equipment and a storage medium, and relates to the technical field of geophysical exploration. The method comprises the following steps: acquiring three-component micro-logging data of a work area, surface wave data of a ground micro-motion monitoring signal and seismic excitation surface wave data; obtaining a surface wave frequency dispersion spectrum of the ground micro-motion monitoring signal and a surface wave frequency dispersion spectrum of the earthquake excitation surface wave data based on the surface wave data of the ground micro-motion monitoring signal and the earthquake excitation surface wave data; automatically picking up the surface wave frequency dispersion spectrum of the ground micro-motion monitoring signal and the surface wave frequency dispersion spectrum of the earthquake excitation surface wave data to obtain a surface wave frequency dispersion curve of the work area; obtaining a near-surface three-dimensional shear wave velocity model of the work area based on the three-component micro-logging data and the surface wave dispersion curve of the work area; and determining the shear wave static correction value of the work area based on the near-surface three-dimensional shear wave velocity model. By adopting the technical scheme provided by the embodiment of the invention, the calculation precision of the transverse wave static correction value of the work area can be improved.

This invention discloses a downscaling optimization method for mountain wind fields based on high-precision simulation, belonging to the field of meteorological numerical simulation technology. The method includes acquiring topographic elevation data and meteorological monitoring data of the target mountain area; identifying abrupt slope changes and constructing a spiral sampling path; calculating aspect variation values ​​to determine the sampling point spacing and generate target sampling data; performing orthogonal decomposition on the sampling data to obtain frequency components; calculating the surface wave coefficient to construct boundary perturbation equations to obtain surface stress distribution; calculating airflow motion state based on surface stress distribution; solving the vorticity equation to obtain vorticity characteristics of the leeward region and calculating airflow motion correction parameters; and performing downscaling iterative calculations with the meteorological monitoring data to generate target area wind field data with a spatial resolution of hundreds of meters. This invention improves the simulation accuracy of local wind fields under complex terrain conditions.

The application discloses the network and information security technical field, specifically a kind of encryption traffic analysis method based on protocol perception state space and end cloud cooperation, sequentially execute bidirectional flow aggregation and multi-view feature adaptive representation, linear state space sequence modeling based on pre-protocol gate, macroscopic feature's residual gate attention purification, the asymmetric trust fusion of perception confusion with temperature annealing, based on heuristic confrontation and joint loss driven's fuse mechanism optimization, based on double defense and limited low-rank fine-tuning's asynchronous cooperative evolution system;Based on this, by mapping the data packet arrival time interval into the input level physical pre-calibration gate, under the premise of completely preserving the state space model extremely low linear complexity and hardware parallel scanning acceleration advantage, the historical feature cross-contamination of cross-session and dormant period message is blocked from source, which greatly improves the representation accuracy of the model in real long sequence network flow.

The application discloses a fault graph structure generation method and system, and the method comprises the following steps: obtaining seismic catalog data of a research area, calculating the triggering probability relationship between seismic events based on a contagious aftershock sequence model; dividing the research area into spatial grids, and performing constraint screening in combination with active fault spatial data, defining the spatial grid unit which simultaneously satisfies the spatial intersection with the active fault and internally contains the seismic event as a fault-constrained grid unit as a node; performing spatial aggregation of the triggering probability relationship according to the fault-constrained grid unit, accumulating the triggering probability between the seismic event pairs belonging to different nodes and satisfying the time sequence, obtaining the triggering intensity between the nodes as the edge weight; and generating the fault graph structure based on the nodes and the directional edges carrying the edge weight. By introducing the event triggering probability into the spatial unit edge weight, the active fault constraint is constructed and applied, and the physical description accuracy of the seismic triggering network to the regional seismic activity structure and the geological interpretation capability are improved.

The invention discloses a precise detection method for gas escape in a spin-coated carbon material curing process, and belongs to the technical field of chemical detection. The method can solve the problem of quantitative detection of gas escaped from the spin-coated carbon material in a high-temperature process, thereby guiding the molecular structure design of the spin-coated carbon material, reducing the film defects, such as micropores, unevenness and insufficient filling, caused by gas release in the process of the spin-coated carbon material, ensuring the strict requirements of high-precision pattern transfer required by the process, and improving the product quality. Therefore, the overall yield is improved, and the reliability of the device is guaranteed.

The application discloses a kind of based on the superdeep layer carbonate rock hillock beach body delineation method of frequency division attribute, the present technical comprehensive application of innovation is used in this method, by the full combination of geological model and seismic prediction method, using the constraint of geological prediction model, reduce the multi-solution of seismic prediction means, solve the limitation of single geology or seismic prediction method, improve the prediction accuracy of hillock beach body plane distribution, can be better used in the exploration and development of hillock beach body reservoir;Traditional method focuses on the conventional attribute delineation in full-band data body, and the prediction accuracy of hillock beach body is limited, and the frequency division attribute based on high-frequency and low-frequency information of seismic data can finely delineate the boundary range of hillock beach body, and also can effectively improve the delineation accuracy of hillock beach body details.

The application discloses a crop environment quantitative evaluation and decision system based on growth stage self-adaptation and relates to the technical field of agricultural wisdom decision-making.The data acquisition module collects original environmental data sequences of multiple types of sensors in the field;the dynamic identification module analyzes the sequences by using a preset crop growth stage identification model, identifies the current growth stage, and outputs a corresponding key environmental parameter weight template;the feature fusion module fuses the original data according to the template to generate a stage-adaptive environmental feature vector;the state evaluation module converts the vector into a growth state quantitative evaluation value through a growth state evaluation model;and the decision trigger judgment module compares the preset dynamic threshold interval corresponding to the growth stage according to the growth state quantitative evaluation value, and generates a decision trigger instruction when the growth state quantitative evaluation value deviates.The application realizes stage self-adaptive intelligence of crop growth monitoring and decision-making, and improves the accuracy of environmental state evaluation and the timeliness of management decision-making.

The application relates to a PM 2.5 Prediction method belongs to the field of air quality prediction. The method comprises the following steps: S1: obtaining air quality data for a period of time, and preprocessing the air quality data to obtain a feature data set; S2: constructing a space-time hybrid model MCMANet, including a time module for capturing the time characteristics of the air quality data; a space module for obtaining the spatial characteristics of the air quality data; and a prediction module for predicting the future PM 2.5 concentration by fusing the time characteristics and the space characteristics; S3: inputting a training set into the MCMANet model for training until the model outputs the most accurate PM 2.5 concentration prediction value; and S4: inputting the preprocessed air quality data at the current moment into the trained MCMANet model to predict the future PM 2.5 concentration. The method can comprehensively improve the accuracy and reliability of PM 2.5 prediction.

This invention relates to a disease evolution prediction system based on cross-modal contrastive learning and counterfactual graph reasoning, belonging to the field of medical information processing technology. It includes: a multimodal heterogeneous data alignment module for aligning multimodal heterogeneous clinical data using cross-modal contrastive learning; a dynamic evolutionary time-series graph construction module for abstracting the complex evolutionary process of chronic diseases into a dynamic graph structure that changes over time; and a counterfactual intervention effect evaluation module that uses a conditional diffusion model as a counterfactual sample generator to construct multiple parallel potential evolutionary paths to evaluate prognostic differences under different treatment strategies. This invention significantly improves the representation accuracy of heterogeneous medical data through cross-modal contrastive learning, effectively compensating for the limitations of missing information in single modalities. This invention not only enhances the robustness of the prediction system in dynamic intervention environments but also endows the model with deep clinical interpretability.

The invention provides a three-dimensional seismic exploration method and system based on multi-source geological data fusion, and the method comprises the steps: carrying out the data standardization of obtained multi-source heterogeneous geological data of a target work area, and generating a standard data set; wherein the multi-source heterogeneous geological data comprises seismological observation data, drilling data and magnetotelluric sounding data; based on the seismic data and the logging data in the standard data set, feature correlation analysis is carried out through a pre-trained deep learning model to obtain a lithology sensitive factor data body, and the lithology sensitive factor data body is used for representing the response confidence coefficient of the seismic data to lithology change; and establishing an objective function of seismic and electromagnetic data joint inversion by taking the lithology sensitive factor data volume as a dynamic constraint, and obtaining comprehensive physical property data by solving the objective function, the comprehensive physical property data being used for constructing a three-dimensional geologic model. The method can improve the precision and reliability of seismic exploration.

The application discloses a kind of fusion multivariate earthwork distribution space interpolation method and device, electronic equipment, it is related to water transport engineering planning technical field, the method includes: based on the topographic surveying and mapping and geological drilling data of canal engineering, determine topographic data, borehole soil layer data and lithology data, adopt analytic hierarchy process to determine multivariate coefficient and the weight parameter of different lithology, based on the weight parameter of different lithology and topographic trend residual variation function is constructed, and spatial interpolation calculation is carried out using machine learning strategy, obtain initial earthwork distribution interpolation result, to construct the stratigraphic interface model corresponding to each stratum, and generate three-dimensional geological model based on all stratigraphic interface model, for each grid element is assigned parameter, output earthwork space distribution chart.The application solves the technical problems that the earthwork calculation mode in the related art has missing geological feature expression, weak adaptability to complex terrain, resulting in large error of output earthwork result.

The invention provides a high-frequency signal processing method for a submarine pipeline SAS monitoring system, and relates to the field of signal processing, and the method comprises the steps: obtaining an original echo signal, converting the original echo signal to a distance frequency domain, carrying out the inversion correction of each layer sound velocity parameter of a layered medium model through the phase characteristics of a submarine pipeline target in echoes, and building a sound velocity recursion relation; calculating one-way propagation time corresponding to different distance frequencies, obtaining an equivalent frequency modulation parameter through numerical difference, and constructing a phase correction value in a distance frequency domain to compensate nonlinear distortion of a distance migration curve caused by sound velocity layering; after distance compression is completed, actual propagation time sequences at different azimuth moments are solved according to a time delay-frequency relation, and a correction coefficient is constructed through azimuth frequency domain group delay analysis to compensate azimuth phase distortion. According to the method, physical matching compensation of signal distortion caused by layered media is realized, and the focusing performance and the geometric fidelity of the SAS image of the submarine pipeline are remarkably improved.

This application relates to a method for predicting the evolution of online public opinion under the intervention of AI agents and online trolls. The method includes: acquiring social media user data; constructing a weighted directed social network containing ordinary netizens, online trolls, and commentary-type AI agents and identifying node types; mapping text content to opinion values, setting the opinion values ​​of online trolls to fixed values, and setting the commentary-type AI agents to generate targeted comment opinion values ​​for different target users; calculating the individual confidence threshold of ordinary netizen nodes, and performing opinion updates based on the extended Deffuant-Weisbuch dynamics equation; iteratively updating until convergence, and outputting the final opinion distribution as the prediction result. This invention, by introducing multi-agent heterogeneous interaction and human intervention, solves the problem of existing models' inadequate characterization of complex public opinion environments, significantly improving the accuracy and modeling completeness of online public opinion evolution prediction.

The invention discloses a carbonate rock crack modeling method based on digital outcrop and geomechanical simulation. According to the method, firstly, a high-precision digital outcrop model is constructed through unmanned aerial vehicle oblique photography, outcrop surface crack parameters are obtained, and crack groups are divided; and determining the ancient stress field characteristics of multiple fracture forming periods by combining filler isotope test and acoustic emission experiment. And further, carrying out geomechanical simulation based on the recovered paleostructure, and quantitatively calculating the density and opening parameters of the crack in each stage by utilizing a two-stage fracture criterion and a strain energy density theory. And then, a deterministic modeling technology and a random modeling technology are fused, a multi-stage and multi-scale discrete fracture network model is constructed, and fluid simulation is performed based on the model so as to analyze the control effect of the fracture on the ancient karst. And finally, through comparison verification with outcrop measured data, the corrected crack-matrix coupling prototype geologic model is output, quantitative prediction of the cracks in the outcrop is realized, and the precision and reliability of the crack prototype model are remarkably improved.

The invention relates to a tobacco production data feature screening technology, in particular to a method and a system for characterizing cigarette end shred falling quantity based on prior process knowledge and a storage medium, and the method comprises the following steps: obtaining cigarette production process data and carrying out dynamic region division; performing multi-dimensional feature extraction on the data after region division to obtain a multi-dimensional feature set; performing edge feature pre-screening according to the multi-dimensional feature set to obtain an effective feature set; obtaining a comprehensive score feature set according to the effective feature set; performing intra-region feature screening on the comprehensive score feature set to obtain an initial feature set and a candidate feature set; and performing improved sequence forward selection fine screening based on the initial feature set and the candidate feature set to obtain an optimal feature subset. According to the characterization method for the tobacco rod end portion shred falling amount, higher characterization precision, higher model robustness, better process interpretability and wider production scene adaptability are achieved.

The invention discloses a remote node dormant state machine control method and system based on message time sequence analysis, and relates to the field of node power consumption management control, and the method comprises the steps: obtaining an arrival timestamp and a message identifier of an inbound message, and carrying out the classified storage of the arrival timestamp and the message identifier according to service types; extracting statistical characteristics based on the adjacent arrival time interval sequence of each category of messages, constructing a message arrival probability distribution model and calculating an arrival risk degree function; carrying out weighted fusion by combining service priorities to obtain a system-level comprehensive risk degree function, and predicting that a next message arrives at an expected moment; and calculating a safety margin in combination with the current energy state of the node, and selecting a target dormant state in the multi-stage dormant state machine. Therefore, dynamic control over the sleep depth and the wake-up time is achieved, and the adaptivity and response coordination of low-power-consumption management of the far-end node are improved.

The invention discloses a lithium battery SOC prediction method fusing online parameter identification and QPSO-AUKF and application, an adaptive unscented Kalman filter (AUKF) algorithm is adopted to construct a prediction model, a QPSO algorithm is added, and particle positions are described through a probability density function. Optimal parameters of a process noise covariance matrix and an observation noise covariance matrix are searched more efficiently in a solution space, so that the adaptability and estimation precision of the filter under different working conditions are improved; when SOC estimation is carried out, FFRLS is adopted to recognize battery parameters in lithium battery modeling in real time, model parameters are recursively updated to obtain optimal parameter vectors changing along with time, the optimal parameter vectors are input into a QPSO-AUKF algorithm, and a closed-loop correction structure is formed in the cycle recursion process of a prediction model. According to the method, the system characteristic change can be accurately captured, and the synchronous updating of the battery parameters and the state is realized, so that the precision and the stability of SOC estimation are remarkably improved.

The application provides a traffic flow prediction method fusing curvature gradient, and relates to the technical field of traffic flow prediction. The method first acquires traffic data and geographic information system data of visible area detectors in a road network; extracts curvature and gradient characteristics of each road section, and constructs a geographic enhanced road network graph structure, wherein the edge weight between nodes is dynamically generated according to the curvature difference and the gradient difference; the traffic data is input into a graph neural network model as node features and graph structure, the model performs spatio-temporal modeling through a graph convolution operator dedicated to curvature and gradient, and explicitly fuses road alignment constraints to capture traffic flow propagation rules; finally, the prediction value of the traffic state of the monitored blind area road section is output, and dynamic prediction is realized. The application effectively solves the problem of inaccurate traffic state perception caused by monitoring blind area under complex terrain, and improves the prediction accuracy and reliability.

The application relates to the technical field of LNG ship containment system detection, in particular to a hydrostatic pressure test tool which comprises a base, a pressure tank and an annular sealing piece. The base is used for bearing a to-be-detected module; the pressure tank is arranged on the base and is detachably connected with the base; one end of the pressure tank is provided with an opening and is sealed by the base; a test cavity for containing test liquid is formed between the base and the pressure tank; the annular sealing piece is sleeved on the outside of the to-be-detected module, the inside of the annular sealing piece is sealingly connected with the edge of the film, the outside of the annular sealing piece extends to the cooperation surface between the base and the pressure tank, and the outside of the annular sealing piece is sealingly abutted with the base and the pressure tank. The application can make the to-be-detected module accept inspection in the state close to the actual installation and loading, so that the comprehensive mechanical properties and structural reliability of the to-be-detected module can be more accurately evaluated.

The invention discloses a bridge crack segmentation method based on hierarchical curvature mask and neural topology evolution, and the method comprises the steps: obtaining a bridge image, executing two-dimensional frequency domain decoupling preprocessing, and separating a low-frequency illumination background through a band elimination filter; calculating a Hessian matrix of the feature tensor to extract a tubular main curvature, generating a layered mask to reserve the fracture skeleton features and executing sparse reconstruction; global priori of the visual basis large model is injected into the backbone network by using a frequency domain low-rank feature adapter, and global context features are extracted in combination with a discretized selective state-space equation; introducing continuous implicit neural representation, and converting discrete pixel segmentation into probability field regression of a continuous coordinate space; the continuous coherence theory in algebraic topology is introduced in the fine tuning stage, and penalty constraint is applied to fracture disconnection errors by calculating the Wasserstein distance of a topological persistence graph. According to the method, discrete quantization errors are effectively overcome, and segmentation connectivity and precision under a complex background are improved.