37 results about "Composite sample" patented technology

The invention relates to a heterogeneous insulation interface parameter testing method and device based on broad-spectrum photoelectric emission, and the method comprises the steps: applying metal electrodes with different work functions to to-be-tested samples of two single-layer materials forming heterogeneous insulation, and obtaining a single-layer material photoelectric emission current spectrum; analyzing the photoelectric emission current spectrums of the single-layer materials to obtain parameters of the two single-layer materials forming heterogeneous insulation; applying metal electrodes with different work functions to a heterogeneous insulation composite sample formed by tightly attaching two single-layer materials forming heterogeneous insulation to obtain a total photoelectric emission current spectrum of the heterogeneous insulation composite sample; separating according to the total photoelectric emission current spectrum of the heterogeneous insulation composite sample to obtain an intermediate interface photoelectric emission current spectrum; and analyzing the photoelectric emission current spectrum of the intermediate interface, and calculating to obtain a real contact potential barrier between the two single-layer materials forming the heterogeneous insulation. Compared with the prior art, high-precision measurement of heterogeneous insulation interface parameters is realized.

The invention discloses a composite material reflectivity spectral information classification method and system based on a PCA-SVM algorithm. The method comprises the following steps: collecting reflectivity spectral data of a composite material sample; preprocessing data to eliminate measurement deviation and unify numerical scale; carrying out dimensionality reduction on the preprocessed high-dimensional spectral data through principal component analysis, and extracting feature components retaining main variance information; based on dimension reduction features, a support vector machine is adopted to construct a multi-label classification model according to a'one-to-other 'strategy; predicting the test sample, and generating a multi-label classification result through probability output and threshold processing; and analyzing the classification performance by using the multi-label evaluation index. The data processing module of the system executes preprocessing, dimension reduction, modeling, prediction and evaluation operations. The method is suitable for lossless identification of multi-component composite samples, both interpretability and identification precision are considered, the performance bottleneck of a traditional method under the conditions of feature overlapping, insufficient samples and the like is effectively overcome, and efficient and accurate classification of the composite materials is achieved.

The invention belongs to the technical field of binding force testing, and particularly relates to a composite current collector binding force testing device and method.The composite current collector binding force testing device comprises a testing rack, a hot-pressing assembly is installed on the surface of the testing rack, and the hot-pressing assembly is used for conducting hot-pressing binding on a composite current collector three-layer composite sample; the composite current collector three-layer composite sample comprises a test main body layer, a bonding layer and a supporting layer which are sequentially stacked, a force measuring assembly is installed on one side of the hot pressing assembly, the force measuring assembly is used for testing the binding force of the composite current collector three-layer composite sample, the force measuring assembly comprises two clamps, and the two clamps are used for clamping a testing main body layer and a supporting layer of the composite current collector three-layer composite sample respectively. The three-layer composite sample of the composite current collector is subjected to hot-pressing lamination through the hot-pressing assembly, so that the bonding layer is firmly bonded with the metal layer of the test main body layer, and when a test is carried out through the force measuring mechanism, the metal layer can be completely separated, and the accuracy of binding force test data is ensured.

The application belongs to the technical field of material detection, and specifically provides a method for determining the sodium expansion rate of an aluminum cathode with a non-standard size under pressure, characterized in that the method comprises the following steps: preparing a composite sample through an aluminum cathode carbon block and a graphite block, and preparing a control sample through the graphite block; electrolyzing the composite sample and the control sample respectively; determining a first length increment curve of the composite sample in an electrolysis process, and determining a second length increment curve of the control sample in the electrolysis process; and determining the sodium expansion rate of the aluminum cathode based on the first length increment curve and the second length increment curve. The technical scheme provided in the application can realize the determination of the sodium expansion rate of an aluminum cathode with a non-standard size.

This invention discloses a multi-physics in-situ testing platform for aramid-synthetic ester composite insulating materials. The method includes: pre-treating an aramid substrate by drying and impregnation to obtain a composite sample; clamping the sample between test electrodes; applying a stable and controllable temperature field and an electric field to the sample sequentially or simultaneously; and measuring the dielectric constant, dielectric loss factor, and conductivity of the sample in situ under stable thermo-electric coupling conditions. The testing platform includes test electrodes for clamping the sample, a temperature control module for applying the temperature field, an electric field loading module for applying and adjusting the electric field, and a data acquisition module for in-situ data collection. By optimizing the sample pre-treatment process and the synchronous loading control of multi-physics fields, this invention achieves high-precision and highly repeatable in-situ testing of the dielectric properties of materials under actual service conditions within a wide temperature and field strength range. It is particularly suitable for evaluating the performance evolution of composite insulating materials under thermo-electric coupling.

The invention relates to a heat treatment regulation and control method for equiaxial alpha volume fraction of a laser deposition state titanium-based composite material, which comprises the following steps of: heating a sample of the laser deposition state titanium-based composite material to a heating final temperature along with a furnace, immediately cooling the sample to a cooling final temperature along with the furnace at a cooling speed of 1-15 DEG C / min, and taking out the sample from a heat treatment furnace after the sample is cooled to the cooling final temperature, thereby obtaining the equiaxial alpha volume fraction of the laser deposition state titanium-based composite material. The invention further provides a laser deposition state titanium-based composite material obtained through the method, the laser deposition state titanium-based composite material is a TiB2 reinforced Ti65 composite material, and compared with the laser deposition state titanium-based composite material, the yield strength at the room temperature is not lower than 850 MPa, and the tensile strength is not changed or slightly improved by 20-50 MPa; and under the condition that the tensile strength is not lower than 1150 MPa, the room-temperature ductility is improved by 100%-150%.

The invention belongs to the technical field of geotechnical engineering testing, particularly relates to a method for determining composite modulus of waste whole-tire reinforced soil, and aims to solve the problem of data distortion in traditional testing. Comprising the following steps: carrying out a uniaxial tensile test on a waste whole tire to obtain a tire tensile modulus for circumferential stress analysis; the method comprises the following steps: preparing a waste whole tire-filler soil composite sample, carrying out an unconfined compression test, arranging a linear variable differential displacement meter in the diameter direction of the tire to monitor the overall diameter change, and pasting a resistance strain gauge on the inner wall of the tire to measure the local circumferential strain; and calibrating the reading of the resistance strain gauge by using the monitoring result of the linear variable differential displacement meter to obtain calibrated circumferential strain data. And on the basis of a circumferential tension theory, equivalent confining pressure increment provided by the tire to the filler soil is calculated for the obtained calibrated circumferential strain data. And substituting the obtained tire equivalent confining pressure increment into the relationship between the modulus of the filler soil and the confining pressure, and calculating and determining the composite modulus of the waste whole tire reinforced soil.

The application belongs to the technical field of computer vision and artificial intelligence, and discloses a target re-identification method based on frequency domain guidance prompt, equipment and medium, the method comprises the following steps: obtaining a target image to be processed; performing frequency domain decomposition on the target image to extract and fuse high-frequency components to obtain a high-frequency detail image; determining a plurality of high-frequency image blocks containing fine-grained texture information in the image according to the high-frequency detail image, and encoding to generate instance tokens; obtaining a visual prototype corresponding to the image identity from a prototype library, and generating a prototype token; embedding the instance token and the prototype token into a text template to generate an image-specific personalized text prompt; finally, constructing a composite image-text sample pair containing difficult positive and negative samples based on the generated personalized text prompt, and training a visual-language pre-training model using the cross-modal contrast loss mined by the composite sample. The application generates a personalized text prompt for each image and effectively improves the performance of target re-identification.

This invention belongs to the technical field of railway tracks, specifically disclosing a method for identifying rail corrugation in rail transit based on the composite characteristics of train vibration and noise. The method includes the following steps: acquisition and processing of train vibration, noise, and related signals; enhancement and fusion of vibration and noise signals; mapping of vibration and noise composite data to corrugation relationships and establishment of a sample set; design and training of a convolutional neural network structure for the vibration and noise composite sample set; and identification of rail corrugation status. This invention uses a one-dimensional convolutional neural network to adaptively extract features from the sample set of vibration fusion data based on vibration and noise fusion data. Simultaneously, a random search method is used to determine the optimal parameters, shortening the sample training and identification time, and meeting the requirements for accuracy and timeliness in rail corrugation detection and monitoring.

The invention provides a non-circular signal spectrum sensing method based on a weighting decision coefficient, and belongs to the technical field of wireless communication, and the method comprises the steps: building an actual value composite signal matrix based on a complex value receiving signal, and calculating a composite sample covariance matrix; calculating a correlation coefficient based on the composite sample covariance matrix, and constructing a decision coefficient vector; estimating a parent mean value and a parent covariance of the decision coefficient vector to form an estimated mean value vector and an estimated covariance matrix; calculating a weight coefficient vector, and carrying out linear weighted fusion on the weight coefficient vector and the decision coefficient vector to construct a statistic; and calculating a sensing threshold value, comparing the sensing threshold value with the statistical magnitude, and determining a spectrum sensing result. According to the method, the composite covariance matrix for describing the second-order statistical characteristics of the non-circular signals is fully utilized, weight fusion is carried out on the decision coefficients with different intensities in combination with the weight coefficients driven by the data, and better sensing performance can be obtained.

This invention belongs to the field of geotechnical engineering testing technology, specifically relating to a method for determining the composite modulus of waste tire reinforced soil, aiming to solve the problem of data distortion in traditional testing. The method includes conducting uniaxial tensile tests on waste tires to obtain the tire tensile modulus for circumferential stress analysis. Waste tire-fill soil composite samples are prepared and subjected to unconfined compression tests. A linear variable differential displacement gauge is placed along the tire diameter direction to monitor the overall diameter change. Resistance strain gauges are attached to the inner wall of the tire to measure local circumferential strain. The readings of the resistance strain gauges are calibrated using the monitoring results of the linear variable differential displacement gauge to obtain calibrated circumferential strain data. Based on circumferential tension theory, the equivalent confining pressure increment provided by the tire to the filler soil is calculated from the obtained calibrated circumferential strain data. The obtained equivalent confining pressure increment is substituted into the relationship between the modulus of the filler soil and the confining pressure to calculate and determine the composite modulus of waste tire reinforced soil.

The embodiment of the invention provides a thermophysical property measurement method and device for a composite phase change material and a storage medium, and relates to the technical field of testing. The device comprises a contrast sample measurement module, a composite sample measurement module, a temperature acquisition module and a data processing module, the control sample measurement module comprises a control sample fiber and a first thermoelectric module; the composite sample measurement module comprises a composite phase change sample fiber and a second thermoelectric module; the first thermoelectric module is used for heating a control sample fiber; the second thermoelectric module is used for heating the composite phase change sample fiber according to the phase change temperature range; the temperature acquisition module is used for acquiring composite temperature change data and contrast temperature change data; and the data processing module is used for generating a phase change thermophysical relationship diagram according to the composite temperature change data and the contrast temperature change data. According to the embodiment of the invention, the thermophysical property of the composite phase change material can be measured, and the measurement accuracy is improved.

Various implementations include a mold for forming composite samples for a pullout test. The mold includes a body having a first surface and a second surface spaced apart from the first surface. The first surface defines one or more slots and one or more channels. Each of the one or more channels has a longitudinal axis. At least one of the one or more channels intersects one of the one or more slots. Various other implementations include a method of forming a composite sample for a pullout test. The method includes providing a mold as described above; disposing a resin within one of the slots; disposing a fiber within at least one of the channels intersecting the one of the slots such that a portion of the fiber is disposed within the one of the slots; and causing the resin to cure.

The invention provides a composite material regenerated by waste and a preparation method thereof, and belongs to the technical field of environment-friendly composites.The preparation method specifically comprises the following steps that firstly, chitosan powder and waste printed circuit board nonmetal powder are mixed and stirred according to the mass ratio of (15-20): (30-60), and mixed powder to be treated is obtained; then at a preset temperature, a certain amount of hydrolyzed silane coupling agent is added into the to-be-treated mixed powder to be stirred, the to-be-treated mixed powder is subjected to surface treatment and preliminary crosslinking through the silane coupling agent, modified powder is obtained, and the mass ratio of the silane coupling agent to the to-be-treated mixed powder is (1.4-1.6): 100; mixing the modified powder, waste polypropylene particles and an antioxidant according to a mass ratio of (30-50): 200: 0.5, and stirring to obtain composite sample powder; and finally, carrying out extrusion granulation on the composite sample powder, and carrying out injection molding to prepare the composite material regenerated from wastes.

The invention belongs to the technical field of data processing, and provides a method and system for rapidly evaluating the carbon sequestration capability of a biochar composite material, and the method specifically comprises the following steps: carrying out field application on marked mineral composite biochar, and carrying out high-frequency real-time monitoring on the composite biochar to obtain the carbon content loss rate; the method comprises the following steps: measuring carbon content loss rate between measurement intervals, acquiring loss deviation according to the carbon content loss rate between the measurement intervals, comparing the loss deviation with a benchmark database to obtain loss mode fitting quality, selecting a reference sample for evaluating the carbon sequestration capability by taking the loss mode fitting quality as an evaluation standard, and outputting evaluation data. The method comprises the following steps: constructing loss mode fitting quality data of a sample in a time sequence dimension, and quickly quantifying the relative stability degree of carbon sequestration capability corresponding to carbon release kinetics in a biochar composite system on the premise of limited monitoring data flow, so that the greenhouse gas emission inhibition effect of the currently researched composite sample can be quantitatively evaluated; and the sampling frequency and the experiment cost required by traditional long-term sample burying analysis are reduced.

This invention belongs to the field of textile printing and dyeing technology, specifically relating to a method for testing the resin fastness of wrinkle-free yarn-dyed pure cotton fabrics for garments. The fabric to be tested is a white-based, dark-striped or dark-colored pure cotton fabric for wrinkle-free garments. The method includes the following steps: Sample preparation: Take the fabric to be tested and sew it into a sample to form a composite sample; Testing process: Place the composite sample in a stainless steel cup, add 20-25g of resin mixture and 80-100 stainless steel beads, and place it in a wash fastness testing machine, running it at 35-45℃ for 40-50 minutes; Rating: Remove the composite sample, rinse with room temperature water, air dry, and use the AATCC EP2 staining grey scale to rate the staining of the multi-fiber interlining. This invention realistically simulates the wrinkle-free resin rolling process, which is closer to actual production than traditional soap washing testing, and the judgment results are more reliable. Early screening before garment processing avoids rework and waste caused by unqualified finished products.

The invention belongs to the technical field of preparation of light metal plates, and particularly relates to a preparation method of a heterogeneous aluminum alloy composite plate. The method comprises the following steps: preheating an Al-Zr plate sample, dipping the Al-Si alloy in a semi-molten state, and cooling to obtain an Al-Zr-Si composite sample; the method comprises the following steps: preheating an Al-Zr-Si composite sample, dipping the Al-Mg-Si alloy in a semi-molten state, and cooling to obtain an Al-Zr-Si-Mg composite sample; and the Al-Zr-Si-Mg composite sample is pretreated and subjected to accumulative ply rolling, and the heterogeneous aluminum alloy composite board is obtained. The prepared heterogeneous aluminum alloy composite board is of a layered gradient heterostructure, the mechanical property layout of'toughness-strength-stronger 'from outside to inside is achieved, the bearing capacity of each layer is fully played, the best matching of strength and plasticity is achieved, and the performance requirements of shells of aerospace and electronic equipment are met.

The invention provides a sample preparation method for a high-temperature confocal microscopy experiment, which comprises the following steps: machining a plurality of sample blanks into a specified size, arranging a plurality of primary samples in parallel, jointly embedding the primary samples into the same sample embedding material to form a composite sample embedding block body, and uniformly grinding and polishing the composite sample embedding block body; and cutting the polished composite mounting block into sheets with specified thickness by using a precision cutting machine, and separating and cleaning the sheet-shaped sample from the mounting material. According to the invention, the preparation process of a plurality of sample blanks is changed from traditional single treatment to batch treatment through embedding integration and then segmentation separation, so that the sample preparation efficiency and consistency are remarkably improved, meanwhile, the safety risk of handheld micro sample operation can be eliminated, and the preparation of a material high-temperature performance test sample is facilitated.

The application discloses a method for measuring the elastic modulus and Poisson's ratio of irregular rock, and belongs to the technical field of geotechnical engineering, and solves the problem that the existing rock mechanics characteristic evaluation technology needs to cut irregular rock samples, and destroys the structure and form of the rock samples. The application combines three-dimensional scanning technology, physical experiments and finite element numerical inversion, does not directly perform end face flattening treatment on original irregular rock samples which is difficult to realize, but uses a pouring material with known mechanical properties to encapsulate and pour the irregular rock samples, forms a composite sample which can be tested, measures the overall response of the composite sample under load, establishes a finite element numerical model which accurately corresponds to the composite sample, and finally inversely calculates the real elastic modulus and Poisson's ratio of the irregular rock through a multi-objective optimization algorithm.

The application relates to a method and device for evaluating the insulation-shielding matching property of a high-voltage alternating-current cable. The method comprises the following steps: obtaining a target insulation sample, and respective shielding samples and respective target composite samples corresponding to the target insulation sample; wherein the respective target composite samples are prepared by respectively compounding the target insulation material and the respective shielding materials; performing first performance tests on the target insulation sample and the respective shielding samples by using a first device to obtain first performance results; wherein the first performance tests at least include thermal expansion performance tests; performing second performance tests on the respective target composite samples by using a second device to obtain second performance results; wherein the second performance tests at least include breakdown performance tests; and selecting a target shielding material matched with the target insulation sample from the respective shielding materials according to the first performance results and the second performance results. The application can improve the accuracy of cable performance evaluation.

The invention belongs to the technical field of computer vision and artificial intelligence, and discloses a target re-identification method and device based on frequency domain guidance prompt and a medium, and the method comprises the steps: obtaining a to-be-processed target image; performing frequency domain decomposition on the target image to extract and fuse high-frequency components to obtain a high-frequency detail image; determining a plurality of high-frequency image blocks containing fine-grained texture information in the image according to the high-frequency detail image, and encoding to generate an instance token; obtaining a visual prototype corresponding to the image identity from a prototype library, and generating a prototype token; embedding the instance token and the prototype token into a text template, and generating a personalized text prompt exclusive to the image; and finally, constructing a composite graph-text sample pair containing samples difficult to be positive and negative based on the generated personalized text prompt, and training a vision-language pre-training model by using cross-modal contrast loss mined by the composite sample. The personalized text prompt is generated for each image, and the target re-identification performance is effectively improved.

The invention belongs to the technical field of environmental monitoring and pollutant analysis and detection, discloses a method for passively collecting and extracting micro-plastics on the surface of an atmospheric leaf of a needle-leaved plant, and aims to solve the problem that no uniform standard is provided for extracting micro-plastics on the surface of atmospheric micro-plastic vegetation in the prior art. At the height of 2.2 m, about 100 g of annual needle leaves are collected from 4-5 needle leaf plants through high branch shears to form a composite sample, and the composite sample is wrapped and sealed with aluminum foil.

This invention discloses a method for detecting the hollowness of hollow glass microspheres based on etching-image analysis, belonging to the field of materials testing technology. It addresses the technical problems of existing technologies being unable to distinguish between hollow and solid glass microspheres with high throughput and accuracy, and being susceptible to imaging drift and surface concavity interference. The method includes: implanting the glass microspheres to be tested into a substrate; solidifying the substrate to obtain a composite sample; chemically etching the composite sample and acquiring a grayscale image containing the morphology of the glass microspheres; identifying substrate areas without particle attachment and extracting their average grayscale as a dynamic background reference; segmenting individual glass microsphere particles after removing background noise; extracting feature parameters of individual glass microsphere particles; and determining hollow, solid, and uncertain particles based on the feature parameters using a preset minimum roundness threshold and a dynamic grayscale difference threshold, and obtaining the hollowness rate.

The application belongs to the technical field of auxiliary tooling and relates to a device for positioning and bonding a resin matrix composite sample reinforcing sheet, comprising a supporting platform, a reinforcing sheet positioning clamp block, a sample positioning clamp block, a length synchronization pin shaft, a width synchronization pin shaft and a pressurizing and temperature control unit, wherein the pressurizing and temperature control unit comprises a connected screw pressurizing table, a heating and temperature control base and an adjustable supporting base; two length synchronization pin shafts are symmetrically arranged on both sides of the pressurizing and temperature control unit along the length direction; the clamp blocks on the same side are connected by the length synchronization pin shaft along the length direction of the clamped sample; the length synchronization pin shaft is slidably connected to one side of the supporting platform; two long hole sliding grooves are prefabricated on the supporting platform, and the pressurizing and temperature control unit is slidably connected in the long hole sliding grooves; the clamp blocks oppositely arranged are connected by the width synchronization pin shaft along the width direction of the clamped sample. The device solves the problem that the existing clamp cannot meet the positioning of test pieces of various specifications and has poor adjustability.

The invention discloses a ballastless track double-concrete tensile property analysis method based on a Brazilian disc technology, and belongs to the field of ballastless track double-concrete mechanical property test.The method comprises the steps that a sample is prepared, specifically, a disc-shaped sample is prepared, and for a ballastless track double-concrete structure, two concrete materials are stacked to be prepared into a composite sample or prepared into single material samples respectively; two ends of the sample are processed into planes, a'platform end face 'design is adopted, the width of a platform is 80% of the diameter of the sample, the flatness of the end face is controlled to be less than or equal to 0.003 mm / m, the parallelism error is less than or equal to 0.005 mm, the surface roughness Ra is less than or equal to 0.8 mu m, and the contact fitting degree with a loading pressing rod is more than or equal to 97%; according to the ballastless track double-concrete tensile property analysis method based on the Brazilian disc technology, the defects of a traditional Brazilian disc technology in a ballastless track double-concrete test are overcome, the accuracy and reliability of a test result are improved, and more effective data support is provided for ballastless track structure design and material optimization.

The invention relates to the technical field of semiconductors, in particular to a method and system for reducing the preparation fragmentation rate of an ultrathin semiconductor test sample. The method comprises the following steps: S1, preparing a sample and a supporting body: obtaining a to-be-tested ultrathin semiconductor chip, processing the to-be-tested ultrathin semiconductor chip to form a to-be-tested sample A, and preparing a supporting body B of which the thickness is greater than that of the sample A; s2, performing layered bonding to construct a composite sample preparation body: preparing the layered composite sample preparation body of an angle base body, a first bonding layer, a support body B, a second bonding layer and a sample A; s3, curing and clamping: cooling and curing the layered composite sample preparation body, and mounting the cured layered composite sample preparation body to a sample preparation fixing device; and S4, grinding the controlled grinding medium and the inclined plane: grinding the sample A under the target grinding posture, so that the sample A forms an inclined plane test area for an expansion resistance test. The composite sample preparation body is constructed through layered bonding, the fragment rate in the sample is reduced, and the success rate of ultrathin semiconductor chip sample preparation is improved.

The utility model relates to a clamp for high-temperature shear strength of a composite material welding interface, and belongs to the technical field of material detection. The clamp comprises an upper clamp body and a lower clamp body, the upper clamp body sequentially comprises an upper clamping rod, an upper platform and an upper vertical block from top to bottom, the cross section of the upper clamping rod and the cross section of the upper platform are circular, the cross section of the upper vertical block is semicircular, and the peripheral face of the upper platform coincides with the peripheral face of the upper vertical block. The lower clamp sequentially comprises a lower vertical block, a lower platform and a lower clamping rod from top to bottom, the sections of the lower clamping rod and the lower platform are circular, the section of the lower vertical block is semicircular, and the peripheral surfaces of the lower platform and the lower vertical block coincide; the arc-shaped bulge directions of the lower vertical block and the upper vertical block are opposite; a through groove is formed in the lower vertical block, and the center line of the through groove is parallel to the horizontal plane. The clamp is simple in structure, accurate in shear strength test result and suitable for a composite sample or a sample with a small size.

The invention relates to a test method for a falling body icebreaking response test of a composite material structure, and belongs to the technical field of ship and ocean engineering composite material structure test. Comprising the following steps: preparation and calibration of an ice body model, preparation and pretreatment of a composite material sample, mounting of a sensor and the sample on a test bed, execution and control of a collision test, and synchronous acquisition and measurement of data. A composite material sample falling body impact ice layer model with a certain balance weight is developed, the failure mode of the ice body and the failure behavior of the composite material in the collision process are studied, and meanwhile the dynamic response of collision between the large-section composite material and the ice body is analyzed.

The invention discloses a polyimide composite material sample extraction structure and an operation method thereof, and relates to the field of sample extraction.The polyimide composite material sample extraction structure comprises a base and a vertical plate installed above the rear end of the base through bolts, a movable platform is arranged at the upper end of the base, a sample box is arranged at the upper end of the movable platform, and a lifting platform is installed at the front end of the vertical plate; a sample extraction mechanism is arranged at the middle end of the lifting platform. According to the polyimide composite material sample extraction structure and the operation method thereof, when the polyimide composite material sample extraction structure is used, a hydraulic rod at the bottom of a base is started, a lifting platform in threaded connection with the surface of the hydraulic rod can be driven along with rotation of a threaded rod, and a limiting rod penetrates through the rear end of the lifting platform; the movable block slides towards the inner side of the lifting platform under the limitation of the cylinders at the front end and the rear end and drives the movable clamping rod and the collecting cover, so that the collecting cover clamps and collects polyimide composite material particles in the sample box, and convenient sampling is realized.

This invention relates to the field of composite sample processing technology in image defect detection, and provides a machine learning-based method for calculating the spatial efficiency of composite samples. The method includes: acquiring the current detection model, multi-source sample data, and corresponding contextual constraint data to construct composite sample spatial units; calculating the boundary gain, drift demand, coverage gap, confidence metric, and redundancy suppression of each spatial unit to obtain the spatial unit efficiency; performing action calculations under constraints of budget, real anchor points, and synthetic injection ratio; performing defect synthesis on the supplementary sample units and updating the confidence metric on the verification units; and performing training injection based on sample confidence and sample constraint relationships to obtain the updated detection model. This invention can perform unified calculations on real samples, weakly labeled samples, and synthetic samples, improving defect recall capability and suppressing false alarm fluctuations under varying process conditions.