45 results about "Hough transform" patented technology

The invention discloses a multi-field part size and appearance defect intelligent detection system, and relates to the field of image analysis. The system comprises an image acquisition module, an image preprocessing module, a feature extraction module, a defect identification and size measurement module, a data processing and analysis module, an automatic control module and a man-machine interaction module. According to the method, CNN and LBP, Hough transform and SIFT algorithms are fused, high-level semantics and bottom-level texture / geometric features are considered, 2304-dimensional fusion feature vectors are formed through feature splicing, the feature extraction integrity of parts in multiple fields is improved, texture detail features can be accurately extracted, geometric shape features can be accurately obtained, and the method is suitable for large-scale popularization and application. Meanwhile, the adaptive feature selection mechanism dynamically optimizes the feature combination according to the detection result, the problem of calculation redundancy is reduced, and the detection precision is ensured while the detection efficiency is improved.

The invention discloses a road traffic detection control method based on a vehicle-mounted camera module, and belongs to the technical field of traffic detection, and the method specifically comprises the steps: collecting road image data in real time through the vehicle-mounted camera module; preprocessing the acquired road image data, and extracting a road area; performing edge extraction on the preprocessed road image data by using an edge detection algorithm to obtain road edge information; according to the road edge information, detecting straight-line segments in the road through Hough transform, and calculating an included angle between adjacent straight-line segments to determine the line type of the road; calculating the distance between the vehicle and the road boundary and the offset of the vehicle relative to the road center line according to the line type of the road in combination with the speed and position information of the vehicle; according to the distance between the vehicle and the road boundary and the offset of the vehicle relative to the road center line, the driving direction and speed of the vehicle are adjusted through a preset control strategy; according to the invention, real-time safety detection of the vehicle in the driving state is realized.

The invention provides a door opening and closing recognition method and system based on AI video analysis, and the method comprises the steps: reading a video file, collecting video data, carrying out the data preprocessing, obtaining a to-be-recognized image, carrying out the door frame recognition of the to-be-recognized image based on edge detection and Hough transform, obtaining a door body region ROI, carrying out the expansion of a detection frame, and obtaining a door body region ROI; key frame extraction is carried out according to the extended detection frame, and motion track characteristics of the door leaf are captured; constructing and training a dynamic state recognition network in combination with a YOLOv11 algorithm and deformable convolution to obtain a target dynamic state recognition model, recognizing motion track features, outputting to obtain an opening and closing motion state of a door, extracting an angle sequence of a recent preset frame number to obtain a plurality of feature values, and inputting the feature values to a time sequence model to obtain a time sequence model; and obtaining a triggering scene which causes the door state to change. According to the invention, generalization identification of opening and closing states of various door types can be realized, continuous angle detection is supported, and the accuracy and efficiency of door state identification are improved.

The invention discloses an AI image recognition-based zero code form automatic generation method, which comprises the following steps of: receiving a sketch or a screenshot provided by a user, and carrying out image Gaussian filtering noise reduction, simplified Canny algorithm edge detection and Otsu method binarization processing on the sketch or the screenshot; using an object detection model to identify form components in the preprocessed image, and outputting bounding box coordinates of each component; detecting a connecting line through Hough transform, judging whether the connecting line is an arrow or not, and extracting a potential relationship between the form components; constructing a dynamic relation graph according to the extracted potential relation, and analyzing a logic relation in the dynamic relation graph by using a graph neural network; a user interaction path is optimized through reinforcement learning, and an interactive form is generated according to the logic relation obtained through analysis and the optimized interaction path; according to the method, the interactive form can be automatically generated from the sketch or the screenshot provided by the user, the form development threshold is lowered, and the form development efficiency is improved.

The invention discloses an underwriting multi-dimensional decision-making method for performing intelligent inspection and evaluation on health risks, and the method comprises the steps: collecting image data of a medical data file, and carrying out the preprocessing of the image data; the preprocessing comprises image enhancement, gray level transformation, Gaussian filtering and Hough transform correction; applying an OCR model based on a neural network, performing structured extraction on the image data according to an underwriting multi-dimensional medical decision rule to obtain structured data in a text form, performing standardization processing on the structured data, and transmitting the processed structured data into the underwriting multi-dimensional medical decision rule; and according to the underwriting multi-dimensional medical decision rule, performing multi-dimensional analysis and evaluation on the standardized structured data based on a preset rule flow, and automatically outputting an underwriting result. The multi-dimensional medical decision rule is adopted to effectively overcome multiple technical difficulties in the underwriting and underwriting fields, manual operation is replaced, the underwriting operation efficiency is greatly improved, the underwriting process is simpler, more convenient and faster, and the underwriting conclusion is more accurate.

The invention provides a desulfurization control method and device for waste incineration, electronic equipment and a storage medium, and relates to the technical field of waste treatment.The method comprises the steps that a combustion multispectral image and the internal average temperature of a waste incineration device in the waste incineration process are obtained; performing region segmentation on the combustion multispectral image according to a deep learning model to obtain a flame region image; performing flame edge contour detection on the flame region image based on improved Hough transform to obtain a flame edge contour; calculating a flame area fluctuation rate according to the flame edge contour, and calculating a temperature change rate according to the internal average temperature; and the adding amount of the desulfurizing agent is dynamically adjusted according to the flame area fluctuation rate and the temperature change rate. Machine vision and a distributed control system are fused and applied to the desulfurization control link of waste incineration, timely and accurate response can be made to complex and changeable working conditions, the use amount of a desulfurizing agent is reduced to the maximum extent, and the operation cost is reduced.

The application provides an image recognition-based online detection method and system for surface defects of ultra-wide and thick plates, and relates to the technical field of image recognition. The method comprises the following steps: synchronously collecting surface images and process state parameters containing final rolling temperature, rolling force distribution, laminar cooling intensity and real-time speed, firstly completing image splicing according to the final rolling temperature and the rolling force distribution to obtain a topological image; then separating a texture channel image by feature decoupling with the laminar cooling intensity as a constraint; subsequently extracting edge features of the texture channel image by Hough transformation, and introducing the rolling force distribution and the real-time speed as constraints for curve fitting to generate topological structure features; finally dynamically adjusting a kernel function by a support vector machine, and analyzing the topological structure features and a preset feature library to obtain a defect detection result. The application can effectively improve the accuracy and adaptability of defect detection under complex working conditions through deep fusion and dynamic guidance of process parameters.

The invention relates to a building group contour automatic generation method based on TLS point cloud, and the method comprises the following steps: S1, obtaining the point cloud data of a building at different site positions, and carrying out the point cloud registration of the point cloud data, and obtaining a complete point cloud model of a building group; s2, carrying out downsampling, noise reduction and cutting on the complete point cloud model to obtain a preprocessed complete point cloud model; s3, carrying out point cloud slicing on the preprocessed complete point cloud model, and intercepting a middle section point cloud; s4, performing point cloud projection on the point cloud of the middle section to obtain a projection two-dimensional plane; s5, dividing the projection two-dimensional plane into grids, then converting the grids into a grey-scale map, and carrying out binarization processing; and S6, performing Gaussian blur, Canny edge detection and Hough transform processing on the binarized grey-scale map to obtain a point cloud contour. Compared with the prior art, the method has the advantages of realizing rapid and accurate building group wall line extraction and the like.

The invention discloses a cable insulation layer thickness automatic identification method and device, and belongs to the technical field of cable sheath detection. The method comprises the following steps: acquiring an image file by using an imread function, extracting a predetermined channel of an image as a grey-scale map, carrying out binarization processing on the predetermined channel through an imbinarize function to obtain a black and white binarized graph, calling hole filling of the binarized graph of which the imfill function is greater than the black and white binarized graph, and establishing a filling circular graph of which the contour is continuous and smooth and the internal hole is filled; measuring a circle center and a cross section in the filling circular ring graph by utilizing a circular Hough transform principle, and calculating the thickness of a circular ring in the filling circular ring graph; and extracting the diameter and thickness of the ring, introducing an ellipse fitting algorithm, calculating the shortest and longest outer diameters of the ring, obtaining an ellipse fitting result, and adjusting the measurement angle to measure the ring thickness of the corresponding angle. According to the method, the standard circular ring detection model is established, the ellipse fitting equation is fitted, irregular shapes occurring in actual production are processed, the detection accuracy is improved, and the adaptability to variation is enhanced.

The application relates to an attention scheme and a strip convolution semantic line detection method based on a deep Hough network, and belongs to the field of image target recognition. The method comprises the following steps: adjusting the picture size of a data set, inputting a pretreated picture into a convolution model to perform feature extraction, attention mechanism and multi-scale feature fusion, and obtaining an output feature result. Different scale features are subjected to Hough transformation and fused feature regression prediction to obtain a recognition result; a strip convolution replaces a previous spatial convolution layer to obtain remote related semantic information of an input image recognition target; a strip-shaped pooling layer and a spatial pooling layer mixed pooling layer are used for pooling of a picture of a forward network; a channel attention module is added to a feature pyramid network to help modeling the importance of picture partitioning, so that the influence of redundant information of an input image on a recognition result is reduced; and a Ghost convolution light-weight network is designed to reduce the calculation amount of the network. Compared with a conventional semantic line detection method, the model parameter amount is smaller, and the recognition accuracy is higher.

A method for identifying girth welds in pipeline detection is disclosed. The method comprises the following steps: a plurality of magnetic flux leakage sensors are evenly arranged in the pipeline, a mileage sensor is set in the circumference of the pipeline, and original magnetic flux leakage data and mileage wheel data of multiple channels are synchronously collected in a sampling manner with equal time intervals; a binary map of pipeline magnetic flux leakage distribution is obtained by median filtering, and the morphological features of weld data are enhanced by asymmetric closing operation; straight line segments in the binary map are extracted by Hough transformation to determine whether there are missed welds. The method adopts a two-dimensional data analysis method to perform a similar image processing method on the data to enhance the data, highlight the morphological features of uneven parts in the magnetic flux leakage sensor data, and capture the girth welds therefrom. The method can determine the girth welds in a two-dimensional space to ensure accuracy, and has low requirements on sensor calibration and computer computing power. In addition, a circumferential periodic extension method is adopted in the judgment to improve the reliability of girth weld identification.

The invention provides a field core box corner extraction and perspective distortion correction method based on multi-scale Retinex fusion. The field core box corner extraction and perspective distortion correction method comprises the following steps: S1, preprocessing a core box image based on anisotropic diffusion and multi-scale Retinex fusion; s2, core box angular point extraction based on topological constraint Shii-Tomasi and graph cut optimization interaction; s3, perspective correction based on thin plate spline transformation and Markov random field fusion; and S4, performing correction result precision verification and closed-loop feedback based on Hough transform straight line fitting and geometric moment invariant fusion. According to the method, accurate extraction of core box angular points and efficient distortion correction of images are realized, and meanwhile, full-dimension precision verification of correction results, adaptive migration of processing parameters of batch images and lightweight field deployment of algorithms are completed.

The invention provides a license structured data extraction method and system, and is applied to the technical field of data processing. According to the method, the format is determined through the first five characters of the first row of the machine reading area and the features of the visual reading area, the unconfigured new format is dynamically recorded, new format configuration and annotation are generated, new format structured data extraction is completed, and a training data set is updated in real time; a three-level positioning algorithm fusion strategy is adopted, a surname and name grouping consistency score, a date data accuracy score, a structured data acquisition proportion and a positioning point deviation value are calculated by taking machine-readable area information as a reference, a preset priority is adopted for sorting, meanwhile, Hough transform and SIFT feature matching are applied to process image deviation, and a positioning processing comprehensive result is generated; taking the verified accurate data of the machine reading area and the visual reading area as reference, automatically correcting and complementing the abnormal certificate data, and generating to-be-detected data; and processing the to-be-detected data based on the target dual-channel OCR model to generate target license structured data.

The present application belongs to the technical field of contact network inspection, and particularly relates to a contact network 1C and 4C comprehensive intelligent inspection trolley, which comprises a trolley body, a data acquisition device and a data processing device; wherein the data acquisition device synchronously acquires point observation information and image observation information along a contact network corridor during the trolley body advancing, so as to jointly constitute a unified corridor reference map with a track center line, a track surface reference plane, a wire reference band and a support calibration point column; the data processing device adopts a domain Hough transform guided by the corridor reference map and a relationship backfill algorithm to identify and locate a contact network positioner, a dropper and a sectional insulator under the constraint of the corridor reference map, and finally generates a positioning landing point and a stable pose of a corresponding target in the corridor reference map. The present application effectively improves the detection precision, stability and processing efficiency, reduces the requirement for manual intervention, and significantly improves the maintenance quality and operation efficiency of the railway contact network.

The invention relates to a method for calibrating a vehicle sensor (5) of a motor vehicle, the method comprising the following steps: determining (S1) sensor data by means of the vehicle sensor (5) at a plurality of measuring times, wherein the motor vehicle moves relative to an object in the surroundings of the motor vehicle; calculating (S2) an object position of the object based on the determined sensor data; calculating (S3) a Hough transform based on the calculated object position; determining (S4) an orientation of the vehicle sensor (5) relative to a travel axis (A1) of the motor vehicle based on the calculated Hough transform; and calibrating (S5) the vehicle sensor (5) based on the determined orientation of the vehicle sensor (5) relative to the travel axis (A1) of the motor vehicle.

Advanced Hough-Based On-Device Document Localization. In an embodiment, lines are detected in an input image of a document. The lines are searched for candidate quadrilaterals. For at least a subset of the found candidate quadrilaterals, a contour score is calculated, and the candidate quadrilaterals are saved or discarded based on their contour scores. For each saved candidate quadrilateral, a contrast score is calculated. A final candidate quadrilateral is selected, based on the combined contour and contrast scores for the saved candidate quadrilaterals, to represent the borders of the document.

The invention discloses a laser welding cutter for intelligent production of a multilayer environment-friendly bag and a control method of the laser welding cutter, and relates to the technical field of laser processing. According to the device, a double-shaft transmission system is formed by the first motor and the first lead screw on the two sides of the workbench and the second motor and the second lead screw in the supporting frame, and high-precision positioning of the laser generator in the XY plane is achieved. A fixed pressing plate on the inner side of the workbench is driven by an electric telescopic rod, flexible clamping of materials is achieved through a silicone rubber elastic buffer layer and anti-skid textures, and deformation is avoided. The control method comprises the steps that a CAD file is imported to generate a cutting path, the cutting path is deviated inwards to generate a welding path, cutting parameters and welding parameters are automatically configured based on a material database, a joint cutting image is collected through an industrial camera, deviation is detected through a Canny operator and Hough transform, and the welding path is corrected. According to the invention, the bottleneck of poor edge quality and low efficiency of traditional machining is broken through, and the automation of the whole process of cutting, detecting, correcting and welding is realized.

The application provides a method for judging the change of stored grain based on image processing. The application uses the local grain depression information of different time periods to compare the depression profile range, thereby avoiding the misjudgment caused by the change of stored grain to a certain extent. The method can achieve the purpose of automatically detecting the stored grain condition regularly. The application comprises a grain loading line detection module and a local grain surface detection module. First, the grain loading line detection module performs image preprocessing on the collected grain warehouse image, and then extracts the grain loading line by using Hough transformation, and cuts the image based on the line. Then, the local grain surface detection module performs edge detection on the cut image, and extracts the profile of the grain surface depression. Finally, the stored grain change detection module compares the extracted depression profile with the previous profile to judge whether the grain loading amount of the grain warehouse changes. The application can improve the accuracy of judging the change of stored grain.

The invention discloses a sea ice condition automatic real-time identification method based on shipborne acceleration measurement, the identification method is realized through an identification system, and the identification system comprises three modules: a signal preprocessing module, a time-frequency feature extraction module and a sea ice condition reasoning module; the method comprises the following steps: 1, converting acceleration data into a time-frequency image through time-frequency analysis, and adopting short-time Fourier transform and Hough transform; 2, time-frequency image feature recognition is carried out to extract key time-frequency features, and a DETR model can be adopted; 3, establishing a mapping relation between acceleration signal time-frequency characteristics and sea ice conditions, wherein machine learning methods such as a support vector machine and the like can be adopted; compared with an existing direct measurement technology, the sea ice condition is inversed in real time by adopting a reverse engineering method, the identification accuracy is improved, the cost is reduced, and assistance can be provided for polar region safety operation and navigation decision making.

The invention discloses a digital processing method and system for Zeeman effect experiment data, and relates to the technical field of atomic physics experiment data processing, and the method comprises the following steps: S1, image reading and preprocessing: obtaining a Zeeman effect interference image, carrying out the self-adaptive zooming according to the image size, converting the Zeeman effect interference image into a gray image, and carrying out the self-adaptive zooming according to the image size; enhancing the contrast ratio through self-adaptive histogram equalization, and removing salt and pepper noise by adopting 3 * 3 median filtering; s2, interference ring identification and measurement: carrying out edge detection by adopting a Sobel operator, detecting an interference ring in the image through Hough transform, screening two maximum concentric circles closest to the center of the image, automatically calculating the diameter of the ring, and converting the diameter of the ring into a physical size; s3, key physical quantity calculation: calculating a wavenumber difference, substituting the wavenumber difference into the optimized formula to calculate an electron charge-to-mass ratio, and performing error propagation analysis. Through automatic image processing and a precise algorithm, manual measurement errors are avoided, and the relative error of the electron charge-to-mass ratio can be reduced to 1.07% or below.

The present invention relates to a method for collecting status information of indicator lights in a traffic control equipment cabinet, including the following steps: collecting and preprocessing an image of the traffic control equipment cabinet; performing a semi-Hough transform on the image to calibrate the coordinates of the indicator lights in the equipment cabinet; obtaining the actual color gamut of each indicator light display, capturing the moment when the indicator light changes, and obtaining the display color of the indicator light.

This information processing method involves causing a computer to execute information processing for extracting a straight line corresponding to a long inspection object from an image captured by using an unmanned moving body. The computer executes: processing for generating a filter-processed image having a line of a specific width emphasized from the captured image through spatial filtering processing based on a first filter having selectivity for a line having a specific width corresponding to the inspection object; and processing for extracting, on the basis of Hough transformation processing on the generated filter-processed image, the straight line corresponding to the inspection object from a plurality of straight lines in the filter-processed image.

This invention discloses a positioning device for tensile testing of fiber-reinforced composite materials under cryogenic conditions, comprising a specimen positioning system, a refrigeration system, and a specimen perpendicularity adjustment system. The refrigeration system includes modules for overall environmental refrigeration, test zone refrigeration, and specimen refrigeration. The specimen positioning system includes a first tensile clamp and a second tensile clamp. The specimen perpendicularity adjustment system includes a suction cup, a multi-degree-of-freedom robotic arm, a gripper, an industrial camera, and an image processing system. The rotation angle calculation includes steps such as binarization processing, edge detection, and Hough transform. After determining the specimen angle and position, the internal telescopic shaft of the tensile clamp extends, pushing the gripper to move and clamp the specimen. Finally, different refrigeration methods are selected according to different experimental requirements to perform tensile testing of the fiber-reinforced composite materials. This invention, through a positioning device for tensile testing of fiber-reinforced composite materials under cryogenic conditions, achieves multi-stage refrigeration of the test environment and automatic specimen replacement and perpendicularity adjustment.

The invention discloses a periscopic camera module detection method and system, a computer and a storage medium, and the method comprises the following steps: S1, loading a detection image of a detected camera module, and carrying out the preprocessing of the detection image; s2, positioning the center of the detection image; s3, extracting an edge line of the detection image; s4, calculating the distance from the center of the detection image to the edge line; and S5, judging whether the distance calculated in the step S4 exceeds a preset error or not, if so, determining that the tested camera module is unqualified, and if not, determining that the tested camera module is qualified. According to the invention, automatic and efficient periscope lens module detection means can be realized through template matching, edge detection, Hough transform, distance calculation and other methods based on computer vision and computer graphics related principles, and the detection efficiency and accuracy are improved.

The invention particularly relates to a method for automatically identifying and correcting a handwritten answer sheet, which relates to the field of image processing and comprises the following steps of: firstly acquiring an image and preprocessing the image, determining the edge and the inclination angle of the answer sheet by means of Hough transform, correcting the edge and the inclination angle of the answer sheet into a standard rectangle and performing interpolation processing, and then dividing regions for histogram equalization. Constructing an XML (Extensive Makeup Language) format template database aiming at the manual annotation of different types of answer sheets, and matching with the preprocessed image; and then segmenting the matched image instance, determining a character segmentation position through vertical and horizontal projection analysis, adopting an improved morphological algorithm for adhered characters, and optimizing a segmentation result in combination with structural features and context information. Characters are recognized, confidence is evaluated, and manual intervention is provided for low confidence. And recombining the characters into a text and verifying, scoring according to semantic similarity, and finally feeding back a correction result to a user in a report form to realize full-process automatic correction.

The present application relates to the technical field of image recognition and industrial vision detection, and discloses an electronic atomizer sensing signal data reading method based on image recognition, which comprises the following steps: collecting panoramic images through an industrial camera array and simultaneously locating LED indicator lights and tester screens by using a double-class YOLO detection network; estimating and compensating the interference of LED halos on screen areas based on a point spread function; unsupervisedly analyzing screen layouts through multi-scale Canny edge detection and Hough transformation; respectively performing OCR recognition and skeleton extraction on screen numerical reading areas and waveform areas; after adaptive brightness normalization and optical crosstalk compensation are performed on LED sub-images, extracting timing code matching sensing signals; and finally cross- verifying double-source data and outputting batch reading results.

The invention relates to a multi-scene robustness sea-sky-line detection method, in particular to a multi-scene robustness sea-sky-line detection method. The objective of the invention is to solve the problems of large use scene limitation and poor robustness of various mainstream sea-sky-line detection algorithms at present. The method comprises the following steps of: 1, detecting an input image; if the image is the RGB image, extracting an R channel image and a B channel image of the RGB image, and independently storing the R channel image and the B channel image; executing the second step; if the image is a grayscale image, executing the step 2; 2, obtaining a low-frequency image IM; 3, obtaining an edge image EM; 4, performing connected domain removal on the edge image EM, and performing expansion on the image after the connected domain removal; 5, obtaining an alternative straight line set and a numerical value of an accumulator HN in Hough transform corresponding to each straight line in the alternative straight line set; and 6, judging the straight lines in the alternative straight line set through an HGS method to obtain a real sea-sky-line. The method is applied to the field of sea-sky-line detection.

The invention discloses a vegetation beach scour profile image recognition and mapping method, and belongs to the technical field of beach profile image processing. The method comprises the following steps: extracting a video frame to obtain a color image data set I0, and converting the color image data set I0 into a grayscale image set I1; aiming at the grayscale image data set I1 and then aiming at the grayscale distribution characteristics of different regions, respectively enhancing region characteristics to obtain a grayscale image set I2; performing edge detection on the cross section of the sandy beach according to pixel features of different partitions of the grayscale image data set I2; firstly, Hough transform processing is carried out on an image set I3 after edge detection, and a section point set I4 is extracted; and performing coordinate mapping on the extracted section point set I4 based on a selected reference point by using a least square method, and connecting and mapping edge points by using a smooth cubic B-spline curve. The processing efficiency is high, and manual intervention is greatly reduced; the method is high in complex scene adaptability, outstanding in anti-interference capability, high in profile extraction precision and good in data continuity.

The invention relates to the technical field of image analysis, in particular to an inorganic mineral casting defect visual detection method and system.The method comprises the following steps that extreme point coordinates are generated through two-channel gray level image non-overlapping scanning, direction angles of adjacent points are calculated through Hough transform to form an offset sequence, and a disturbance unit is marked through two-channel direction angle offset operation; connected domain analysis is combined with path length and disturbance quantity to determine a defect area, and principal axis equidistant sampling is carried out to construct a gradient spectrogram matching frequency fluctuation output defect type and coordinate. According to the method, the microscopic texture capturing capability is enhanced through two-channel gray level image non-overlapping scanning, structure distortion is quantized through a Hough transform direction offset sequence, illumination interference misjudgment is reduced through two-channel direction angle synchronous operation, and noise interference is restrained through connected domain two-index constraint judgment. The principal axis equidistant sampling is combined with gradient spectrogram frequency matching to realize accurate defect classification, and the detection precision and the anti-interference performance are improved by spatial arrangement feature quantization and multi-dimensional feature fusion.

The invention provides a main cable erection quality defect online monitoring method and system based on image analysis, and belongs to the technical field of suspension bridge construction quality control. Carrying out defect target positioning on the image with the defect in the preprocessed image by utilizing a YOLO deep learning model, then determining a twisting angle in the image with the defect by utilizing Hough transform, determining a surface damage area in the image with the defect by utilizing a region growing algorithm, judging a twisting condition according to the twisting angle, and determining a surface damage area in the image with the defect. The surface damage condition is judged according to the surface damage area, so that online quality monitoring is carried out in the cable strand erecting process; the technical problems that in the prior art, quality inspection is conducted in a manual inspection mode, the defect recognition error is large, and efficiency is low are solved.