51 results about "Outlier elimination" patented technology

The invention belongs to the technical field of airport clearance safety, and discloses an airport clearance ultrahigh ground feature extraction method fusing LiDAR point cloud and a panoramic segmentation model, and the method comprises the steps: collecting point cloud data in an airport clearance protection region through employing an unmanned plane airborne LiDAR, and carrying out the noise filtering, outlier elimination and datum point-based elevation standardization; constructing a data set with semantic tags and instance numbers, inputting the data set into a SuperCluster panoramic segmentation model, sequentially executing feature extraction, adjacency relation construction, edge weight distribution, local supervision and super-point aggregation, and outputting a segmentation result with semantics and instances; extracting a highest point in each instance point set, and performing spatial comparison with an inner horizontal plane, an approach plane and a transition plane limiting plane established according to an ICAO standard to obtain an instance category, a highest point coordinate, an intrusion depth and a spatial relationship; and performing three-dimensional visual display on the judgment result through the clearance supervision platform.

The invention relates to the technical field of agricultural irrigation monitoring, and discloses a multi-source sensing-based irrigation water utilization coefficient intelligent monitoring method and system, and the method comprises the steps: collecting multi-source data through unmanned plane spectrum sensing, a ground Internet of Things sensor and a crop growth model, carrying out the time-space alignment, abnormal value elimination and standardization processing of the multi-source data, and carrying out the monitoring of the utilization coefficient of irrigation water. Constructing a data set; constructing a three-dimensional digital twinborn model of the irrigation area, and simulating hydrological processes under different irrigation strategies to generate a virtual irrigation water utilization coefficient reference value; performing multi-source feature fusion on the data in the data set through a CNN-LSTM-RF model to obtain target feature data; inputting the target feature data into a PPO algorithm, dynamically optimizing irrigation parameters, maximizing an actual irrigation water utilization coefficient, and outputting an optimal irrigation strategy; irrigation is conducted based on the optimal irrigation strategy, the irrigation water utilization coefficient is monitored in real time, and when the irrigation water utilization coefficient is lower than a set threshold value, early warning information is automatically sent out; according to the invention, irrigation water utilization coefficient monitoring and management are realized.

The invention provides a remote sensing building group depth detection and change intelligent evaluation method for urban planning, and the method comprises the steps: carrying out the multi-modal data collection of a target urban region, and carrying out the standardization processing, abnormal value elimination, multi-modal alignment and interference suppression processing, and obtaining a multi-modal data set; constructing a multi-branch hybrid network, and performing triplet feature extraction on the multi-modal data set through the multi-branch hybrid network to obtain a building group fusion feature map; performing change area detection and change type classification on the two building group fusion feature maps with different time phases to obtain building group change information; and constructing a dynamic evaluation index based on the urban planning core demand, and performing building group change evaluation in combination with the building group change information. According to the method, through multi-source and multi-modal data acquisition and processing, multi-branch hybrid network triplet feature extraction and change detection and dynamic evaluation index application, urban planning core requirements are adapted, and building group detection precision and change evaluation effectiveness are improved.

The invention discloses a photovoltaic power prediction method based on improved empirical mode decomposition and an optimized long short-term memory network, and the method comprises the steps: firstly carrying out the preprocessing of abnormal value elimination, missing value filling, normalization and the like of photovoltaic power and related meteorological data, and improving the data quality; then, an improved empirical mode decomposition (EE-ANEMD) algorithm is adopted to decompose the preprocessed power sequence into a multi-scale intrinsic mode function component and a residual term, and high-frequency noise, intermediate-frequency fluctuation and a low-frequency trend are effectively separated; global optimization is carried out on the hidden layer unit number, the initial learning rate and the maximum number of training times of the LSTM network through an improved sparrow search algorithm (ISSA), finally, the optimized LSTM is utilized to carry out training prediction on each component, and results are fused and subjected to reverse normalization to obtain a final value. Experiments show that the test set RMSE of the method is reduced compared with that of a single LSTM, the mid-term prediction precision is remarkably improved, and reliable technical support is provided for power system dispatching, new energy consumption planning and photovoltaic power station operation and maintenance.

The invention relates to the technical field of computer vision and automatic control, and discloses an automatic checking method for the number of stacked sheet materials, which comprises the following steps: identifying a standard number table image to obtain a platform standard identifier and a material standard number; performing visual identification processing on the field material image, identifying a platform identification target and all material targets corresponding to the platform identification target, and obtaining position information of the material targets; identifying the platform identification target to obtain a platform analysis identifier; and outlier elimination processing and gap filling processing are carried out on the material target according to the position information. And finally obtaining a platform analysis identifier and a material analysis quantity. And matching the platform analysis identifier with the platform standard identifier, comparing the material analysis quantity corresponding to the matched platform with the material standard quantity, and outputting a verification conclusion according to a comparison result. Through the arrangement, accurate detection of the material quantity is realized, and the accuracy, the efficiency and the automation level of warehouse checking work are improved.

The application provides a network load balancing algorithm based on a Xinyuan platform and a Xinyuan terminal, and relates to the technical field of Xinyuan, which comprises the following steps: collecting real-time traffic data from multiple ports; performing normalization, outlier elimination and protocol field standardization preprocessing on the collected data; performing multi-dimensional protocol fingerprint feature extraction on the traffic of each port based on a protocol analysis algorithm; modeling the historical and real-time traffic of each port using a time series neural network, and outputting traffic prediction values and abnormal traffic probability evaluation in a future short period; dynamically generating a multi-port distribution weight factor based on the protocol characteristics and traffic change trend of each port; inputting the prediction results and real-time load state into a reinforcement learning scheduling agent; and immediately performing feedforward load distribution adjustment when abnormal traffic trend is detected. The method can improve the data processing efficiency, realize the intelligence of load scheduling, ensure the resource allocation priority of the ports of the Xinyuan terminal, and guarantee service continuity.

The application discloses an AI model training optimization method and system, relates to the technical field of model training optimization, and comprises the following steps: collecting real-time running data in an AI model training process, and generating a multidimensional real-time running data matrix; performing outlier elimination on the real-time running data matrix, and generating a standardized running data stream; extracting characteristic parameters and coding into a training state characteristic vector; performing multidimensional evaluation on the training state characteristic vector, and generating an original optimization score; generating a comprehensive optimization suggestion; generating hierarchical optimization content, and converting into a personalized optimization report; and forming a training process self-adaptive adjustment cycle based on optimization result feedback. The application solves the problems of low optimization efficiency and difficulty in accurate positioning caused by manual tuning depending on artificial experience, improves resource utilization and optimization accuracy of the AI model training, and realizes intelligent dynamic adjustment of the training process through a closed-loop self-adaptive control mechanism.

The invention discloses a smog methane composite detection method and device, relates to the technical field of smog methane detection, and solves the technical problems of weak anti-interference capability in a data processing link and lack of a multi-parameter compensation and calibration system. The effective acquisition range is dynamically calculated in combination with a factory calibration interval, the signal amplification factor is adaptively adjusted by using a digital potentiometer, the electric signal strength is ensured to be stabilized in the optimal interval, the influence of laser loss and environment temperature on basic data is eliminated from the source, and the filtering coefficient is dynamically adjusted in combination with the signal-to-noise ratio by using a layered filtering strategy. Random noise is filtered out, and key features of an absorption peak are reserved; through segmented weighted polynomial fitting, accurate extraction of core parameters of the absorption peak is realized; and a unified multi-round outlier elimination mechanism eliminates the interference of abnormal data on the methane concentration, the absorption peak position and the smoke sensing value, and improves the credibility of the data.

The invention relates to the technical field of energy conservation and environmental protection, in particular to an AI modeling-based energy station efficient operation system and an operation method thereof, and the method comprises the following steps: collecting equipment operation data at an energy station site through a multi-source sensor, and uploading the equipment operation data to a cloud through an edge gateway; performing time synchronization, abnormal value elimination, missing value processing and standardization processing on the acquired operation data; carrying out working condition clustering on the operation state of the energy station based on the historical preprocessing data, and establishing a corresponding energy consumption baseline; and performing time sequence prediction on the energy station load and the equipment state based on the real-time data and the historical data, and identifying energy consumption abnormity. According to the invention, the accuracy of energy consumption assessment and abnormity identification can be effectively improved, the risk of sudden failure is reduced, invalid start and stop of equipment and energy waste are reduced, the energy utilization efficiency and the equipment operation reliability are improved, the linkage of operation management and operation and maintenance disposal is realized, and the intelligent operation level and comprehensive operation benefits of the energy station are integrally improved.

The invention discloses an unmanned aerial vehicle multi-source fusion navigation positioning method independent of satellite signals, and belongs to the technical field of unmanned aerial vehicle docking navigation, and the method comprises the following steps: 1, building a fusion navigation positioning system; 2, performing time registration and outlier elimination; 3, solving an information distribution coefficient; 4, time-variable measurement noise estimation is carried out; 5, updating time and measurement; step 6, fusion and differential feedback of the main filter; according to the scheme, the problems of insufficient sensor redundancy, rigid multi-source fusion navigation algorithm information distribution and fault isolation mechanism and the like of the existing docking navigation positioning system are solved, so that the navigation positioning precision and the anti-interference capability of the unmanned aerial vehicle are improved.

The application belongs to the technical field of three-dimensional reconstruction and intelligent agriculture, and particularly relates to a three-dimensional Gaussian multi-machine collaborative mapping and optimization method and system for complex agricultural and forestry scenes. In view of the problem of low reconstruction accuracy caused by inconsistent exposure, dynamic disturbance of vegetation and sparse multi-machine view angles in complex agricultural and forestry scenes, the application sequentially carries out multi-machine data preprocessing, spatiotemporal outlier elimination, intermediate frame generation based on FILM and DepthAnything, and 3DGS joint optimization. Dynamic masks are generated through spatiotemporal residual double modal judgment to block dynamic interference at the level of loss function; sparse view angles are completed by using virtual intermediate frames to expand the observation data set. The application can effectively eliminate dynamic artifacts and cavities, realize high-precision three-dimensional reconstruction of complex agricultural and forestry scenes, and be directly applied to digital operation scenes such as vegetation monitoring and automatic path planning of forest land and orchards.

The present application belongs to the technical field of data processing, and particularly relates to a reinforced concrete support construction quality detection method based on data processing. The method comprises the following steps: acquiring raw material proportioning, mixing, pouring and vibrating, and curing environment multi-source real-time data of the whole construction process; performing improved DBSCAN clustering preprocessing (missing value filling, outlier elimination, and standardization); constructing a dynamic correlation model by using a neural network, and optimizing hyperparameters by using an improved starling optimization algorithm; calculating influence coefficients and quality correlation degrees of each link; determining the quality by comparing the quality correlation degrees with a threshold value, and iteratively updating the model until the construction is completed. The present application realizes real-time detection throughout the whole cycle, reduces manual deviation, avoids quality defect solidification, reduces repair cost, and improves the reliability of support structure construction.

The application discloses a UAV multi-modal image registration method and device, medium and product, relates to the technical field of remote sensing image processing and computer vision, and comprises the following steps: calculating the normalized mutual information value of each wave band of the image to be registered and a reference image, selecting a wave band with the maximum value as a reference wave band, solving a global affine transformation matrix by using an enhanced correlation coefficient algorithm, performing global geometric coarse correction, and obtaining the image to be registered after coarse registration; calculating the local principal direction of the image to be registered after coarse registration and the reference image, constructing a Gaussian derivative filter, extracting structural features based on the turning theorem, generating respective structural feature maps and dividing the structural feature maps into regular grids, performing grid-by-grid template matching to obtain a sparse displacement field, performing outlier elimination and interpolation smoothing to generate a dense linear deformation field, performing nonlinear geometric correction and applying the dense linear deformation field to all wave bands of the image to be registered after coarse registration, and obtaining a final registration result. The application realizes high-precision registration of multi-modal images.

The invention relates to the technical field of building energy saving and energy management, in particular to a building park energy-saving potential assessment method and system based on machine learning. During use, a multi-source energy data acquisition module is subdivided into building, floor and equipment levels to acquire energy consumption data, key factors such as meteorology, people flow and equipment working conditions are fused, and the energy-saving potential of a building park is evaluated. A three-principle-based abnormal value elimination scheme and an adaptive data distribution standardization algorithm effectively solve the problems of single data dimension and insufficient continuity, provide high-quality data support for evaluation, extract cross features of temperature and air-conditioning load, people flow and equipment power utilization and the like through feature engineering, and improve the evaluation accuracy. By combining the first sub-model to capture a hybrid model of long and short term dependence and the second sub-model to correct deviation, the problem that a traditional model cannot process multi-factor interaction influence is solved, the prediction precision is improved, and reliable future energy consumption data is provided for energy-saving potential evaluation.

The invention provides a blast furnace molten iron silicon content prediction method based on an improved grey goose algorithm, and relates to the technical field of intelligent prediction, and the method comprises the following steps: collecting multi-dimensional process parameters in a blast furnace operation process, and synchronously recording corresponding silicon content data; performing abnormal value elimination and normalization processing on original parameter data, dividing a training set and a test set, constructing an initial prediction model of the BP neural network, optimizing training parameters of the BP neural network by using an improved grey goose algorithm, and dynamically adjusting step parameters in the algorithm along with the distance between an individual and a global optimal solution. Fusing a velocity field constructed based on a local curvature factor and a local density factor, updating an individual position in a global search stage, reconstructing a final prediction model according to an optimization result, and performing convergence training by using a training set; and finally inputting the test set into the prediction model, and outputting a blast furnace molten iron silicon content prediction result. The method is suitable for blast furnace molten iron component control and has the advantages of being stable in optimization, small in error and high in real-time performance.

The invention provides a frame synchronization and light intensity sampling abnormal value elimination method based on a DLP projection system, and the method comprises the steps: obtaining a to-be-projected original image, extracting a structural feature vector, and calculating the structural prediction light intensity of the original image according to the structural feature vector; collecting at least one light intensity sampling value in a fixed time window after the original image is projected to form a light intensity sequence; calculating the average light intensity of the light intensity sequence, combining the structure prediction light intensity and the structure feature vector, introducing a structure response correction term and a structure adaptability tolerance term, and judging whether the sampling of the current frame is successfully synchronized with the pattern projection or not; if the sampling synchronization of the current frame is judged to be successful, constructing an abnormal value elimination threshold value, and performing abnormal value identification and elimination on the light intensity sequence; calculating the variance based on the residual sampling sequence after the abnormal values are removed; and carrying out weighted fusion on the arithmetic mean value and the median of the residual sequence, and finally generating a light intensity output value and submitting the light intensity output value to an image reconstruction module.

The present application relates to the technical field of building energy saving and energy management, in particular to a building park energy saving potential evaluation method and system based on machine learning, in use, the multi-source energy data acquisition module is subdivided to building, floor, equipment level to collect energy consumption data, and key factors such as meteorology, passenger flow and equipment working condition are fused, based on the 3-principle outlier elimination scheme and the standardized algorithm adapting to data distribution, the problem of single data dimension and insufficient continuity is effectively solved, high-quality data support is provided for evaluation, at the same time, cross features such as temperature and air conditioning load, passenger flow and equipment power consumption are extracted through feature engineering, combined with the hybrid model of the first sub-model capturing long and short-term dependence and the second sub-model correcting deviation, the problem that the traditional model cannot handle multi-factor interaction is solved, the prediction accuracy is improved, and reliable future energy consumption data is provided for energy saving potential evaluation.

The invention relates to a Bluetooth AOA signal outlier elimination method and system based on deep learning, and the method comprises the steps: initializing system configuration, and cooperatively collecting Bluetooth signals through employing a plurality of Bluetooth base stations; acquiring IQ data in a plurality of Bluetooth base stations, and performing data preprocessing; according to the IQ data, performing calculation to obtain a preliminary positioning result of the Bluetooth terminal; constructing an input data set of the deep learning model and designing a network structure of the deep learning model; training a deep learning model according to the input data set; the training precision of the deep learning model is judged, if the training precision meets the requirement, IQ data collected in real time is input into the trained deep learning model, and a real-time positioning result of the Bluetooth terminal is obtained; otherwise, returning to adjust the network structure; and optimizing the real-time positioning result by using a Kalman filtering algorithm to obtain a final positioning result. Compared with the prior art, the method has the advantages of high positioning accuracy and strong robustness.

The application relates to a fuel cell data outlier elimination method, equipment and storage medium, the method comprises the following steps: acquiring fuel cell data comprising a plurality of samples; mapping each sample as a data point, calculating the average distance and variance between the current data point and K adjacent data points for each data point to obtain the corresponding z-score, eliminating the sample corresponding to the data point with the z-score greater than the threshold value sigma from the fuel cell data; for each sample in the eliminated fuel cell data, fitting the variable value based on the variable associated with the fuel cell voltage, calculating the fitness function value of the sample by using the ant colony algorithm, and calculating the fitting degree index value of the polarization curve corresponding to the sample with the highest fitness function value; obtaining the values of the hyperparameters K and sigma corresponding to the maximum fitting degree index value through grid search to obtain the fuel cell data after eliminating the samples under the corresponding values. The application can effectively eliminate the outliers of the fuel cell data.

The invention belongs to the technical field of ship environment simulation, and particularly relates to a digital twinning processing method for the internal environment of a ship. The digital twinning effect of the internal environment of the ship is optimized through multiple optimization processing modes, and the optimization processing modes comprise feature data rapid matching in combination with a traditional point cloud data acquisition scheme, point cloud data outlier elimination and point cloud smoothing processing, multi-level point cloud bounding box feature point cloud extraction and dynamic object database storage optimization. According to the dynamic monitoring optimization processing of the internal environment of the ship based on the target detection technology, the optimization processing of the variable and complex environment in the ship is realized, so that excessive resource consumption and operation process are avoided.

The present application relates to radar data processing technical field, specifically to a kind of radar layer position tracking method of fusing confidence clustering and wavelet energy discrimination, comprising the following steps: S1, the gray processing is carried out to each frame radar image and the candidate layer position point of the edge gradient greater than threshold is extracted using Sobel operator;S2, detect and eliminate outlier, false edge and low confidence point;S3, according to confidence weighted clustering to generate structure point;S4, optimal trajectory is spliced using DTW algorithm to form layer position path;S5, calculate wavelet energy variation rate to correct deviated path;S6, curve fitting and smoothing to generate final layer position trajectory.The present application, by outlier elimination, confidence weighted clustering, dynamic path planning and wavelet energy auxiliary correction mechanism, improves the accuracy, continuity and robustness under complex structure region of radar layer position tracking.

The invention relates to a method for detecting and processing gravity satellite inter-satellite distance variable rate data gross errors. The method comprises the following steps: firstly, carrying out preliminary abnormal value elimination by adopting a Z-fraction method with additional neighborhood mean value constraint, then dividing a long time sequence after preliminary gross error elimination into a daily short time sequence by day, carrying out grid division on the daily short time sequence, calculating a median feature value and a mean value feature value of each grid of the daily short time sequence, and calculating a mean value feature value of each grid of the daily short time sequence; the median characteristic values of the same grid on different days form a median characteristic value time sequence, and the median characteristic value time sequence is fitted to obtain a median characteristic value time sequence fitting value; finally, calculating the residual error of the time sequence fitting value of the mean value characteristic value and the median characteristic value of each grid, if the residual error is greater than or equal to a threshold value, determining that gross error exists in the grid, and if the residual error is less than the threshold value, determining that gross error does not exist in the grid; and if a gross error is suspected to exist, further calculating a residual error of each epoch data in the grid and a median characteristic value time sequence fitting value, and if the residual error of a certain epoch data is greater than or equal to a threshold value, determining that the epoch data is the gross error and removing the epoch data. The trend modeling process and the residual analysis process are closely combined with the spatial-temporal correlation of data, so that the evolution process of anomalies can be described more reasonably, and the physical rationality and method robustness of overall gross error detection are improved.

The application provides a wind turbine data preprocessing method and device, terminal equipment and medium, comprising: in response to a diagnosis instruction of a wind turbine fault diagnosis model, obtaining target operation data associated with the diagnosis instruction; inputting CMS data into a pre-trained integrated convolutional neural network model for anomaly detection classification, identifying that the CMS data belongs to normal signals or abnormal types; performing corresponding data repair processing on the CMS data according to the identified abnormal type to obtain repaired CMS data; and performing outlier elimination and missing value filling on SCADA data to obtain repaired SCADA data. The application can increase the effective sample size of the downstream fault diagnosis model, thereby improving the accuracy of fault diagnosis.

The invention discloses a Transform encoder model generation method for a terahertz metamaterial sensor and the sensor, and the method comprises the steps: enabling obtained training data to generate a plurality of groups of data samples through electromagnetic simulation software; carrying out abnormal value elimination and standardization processing on the multiple groups of data samples to obtain a training data set; the method comprises the following steps: constructing a Transform encoder model; dividing the training data set into a training set and a test set according to a preset proportion; based on the training set, a mean square error is used as a loss function, and a Transform encoder model is optimized through a back propagation algorithm; and based on the mean square error on the test set, determining the dimensionality of the feature vector output by the input embedding layer, the number of subspaces of the multi-head self-attention layer, the learning rate and the optimal value of the batch size, and obtaining an optimized Transformer encoder model for terahertz metamaterial sensor design. The method is used for improving the accuracy of a model prediction result, so that the performance of the sensor is improved.

The invention relates to the technical field of ultrasonic detection, in particular to a steel member welding correction crack detection method and system based on ultrasonic imaging. The method comprises the following steps: acquiring an original ultrasonic signal and performing frequency domain decomposition to obtain frequency domain characteristic data; performing frequency extraction, threshold comparison and de-noising processing on the frequency domain characteristic data to obtain de-noised ultrasonic signals; extracting a waveform peak value based on the de-noised ultrasonic signal to obtain a waveform peak value sequence, and performing defect position calculation on the waveform peak value sequence to obtain a defect area coordinate; performing signal interception based on the defect area coordinates to obtain local echo data; performing gradient calculation and peak value extraction according to the local echo data to obtain a crack coordinate position, and performing interpolation based on the crack coordinate position to obtain a crack boundary point set; and removing abnormal values based on the crack boundary point set to obtain stable support nodes, and calculating a closed boundary based on the stable support nodes to obtain a final crack position.

The invention discloses a flood prediction system fusing space-time convolution and a gated memory network. The flood prediction system comprises a data acquisition unit, a preprocessing unit, a feature extraction unit, a fusion calculation unit and a prediction output unit, the data acquisition unit collects watershed multi-source original data, the preprocessing unit performs abnormal value elimination and standardization, the feature extraction unit extracts terrain, meteorological time-space and climate scene features, the fusion calculation unit performs time-space coupling weight calculation, dynamic gating memory updating and time-space convolution fusion, and the prediction output unit outputs a flood prediction result. The system solves the problem that spatial-temporal characteristics and hydrological mechanisms are disjointed in the prior art, and accurate flood prediction is achieved.

The application discloses an aviation part round hole and end face cooperative detection method, device, equipment and medium, and relates to the aviation part detection field.The method comprises the following steps: performing outlier elimination on the surface three-dimensional point cloud data of the target aviation part by using the 3σ criterion, and obtaining a round hole effective point set and an end face effective point set; based on a double-thread parallel computing architecture, a first thread is used to perform iterative optimization on the round hole effective point set based on the linear least square method and the 3σ criterion, a round hole fitting result is obtained, a second thread is used to extract an end face plane normal vector based on the end face effective point set and the minimum eigenvalue method of the covariance matrix, and an end face fitting result is obtained; based on the fitting result, a normal error component in a three-dimensional round center vector is eliminated by using a vector projection method, so that the projection round center distance and the eccentricity of the nested round hole in the end face plane are calculated, and the concentricity of the nested round hole is evaluated.The application can realize the aviation part round hole and end face cooperative detection in a non-damage, accurate and efficient manner.

The invention discloses a workpiece three-dimensional model construction and processing method based on binocular vision detection data, and the method comprises the steps: firstly carrying out the multi-angle synchronous scanning of a workpiece through employing a combination of a binocular camera and a line laser, and obtaining the multi-view three-dimensional point cloud data of the surface of the workpiece; and constructing a density sensing type dynamic registration framework, and performing distance weighted fusion and outlier elimination on the registered multi-view point clouds in an overlapping region to generate a complete workpiece point cloud. A feature retention type point cloud direct modeling strategy is adopted, normal vector optimization is performed on a down-sampling point cloud, key feature marks are manufactured, a lightweight three-dimensional model is generated through color coding rendering, and a large amount of calculation of traditional surface reconstruction is avoided. And finally, based on model coordinate system calibration and point cloud fitting optimization, extracting functional characteristic parameters of edge straightness, datum plane flatness and hole system position degree, and completing workpiece quality evaluation. According to the method, the point cloud processing speed and the key feature analysis precision are improved, and the real-time requirement of industrial online detection is met.

The application discloses a point cloud outlier elimination system for line structured light scanning, and the grayscale image acquisition module collects multiple images of a measured object frame by frame, carries out filtering and denoising, image normalization and image grayscale preprocessing on each image of the measured object frame by frame, and obtains corresponding multiple two-dimensional grayscale images; the outlier elimination module converts each two-dimensional grayscale image into a binary image through a threshold filtering method, and eliminates a light spot area with a pixel less than a pixel threshold; the outlier elimination module eliminates a light spot area profile with a grayscale average value less than a grayscale threshold; the outlier elimination module calculates a ratio of a profile area of each light spot area to a convex point number of the light spot area, and eliminates two light spot area profiles with the maximum ratio. The application can solve the technical problem of point cloud outliers caused by strong reflection, structural reflection and error recognition of laser stripes in three-dimensional reconstruction of a reflection area in the prior art.

The application discloses an evaluation method and system based on child growth hormone secretion data, and relates to the technical field of medical data processing.The method determines candidate points through robust outlier preliminary screening by acquiring a growth hormone concentration time sequence; extracts a concentration subsequence of a neighborhood window of the candidate points, calculates a dynamic time warping distance of the concentration subsequence from a preset pulse sequence by using a dynamic time warping algorithm, and converts the dynamic time warping distance into a pulse score; performs candidate point reservation, elimination or boundary undetermined point marking according to the score, and generates a first purified concentration sequence from the remaining points; obtains a first fitted concentration sequence by deconvolution reconstruction, and performs LB test on a residual sequence; if significant, performs flip verification based on a residual sum of squares improvement range on the boundary undetermined points in descending order of the score, obtains a third purified concentration sequence, and outputs an evaluation report.The application solves the problems of outlier elimination error and deconvolution system deviation through pulse morphological structure perception and residual feedback verification.