374results about How to "The test result is accurate" patented technology

The invention relates to the field of artificial intelligence, in particular to a fiberboard quality classification method based on machine vision. The method includes: acquiring a fiberboard surface grayscale image; performing sliding window processing on the grey-scale map on the surface of the fiberboard to obtain a plurality of sliding window areas; calculating the smoothness of a central pixel point of the sliding window; taking the pixel point corresponding to the minimum smoothness as a target pixel point, and obtaining a plurality of target connected domains; obtaining gradient amplitudes of pixel points in each target connected domain to construct a corresponding gradient histogram, calculating the probability that each target connected domain is a defect area, and determining all defect areas; calculating the probability of each gray value as a standard gray value, and determining the standard gray value; and calculating the quality coefficient of the fiberboard according to the gray average value of each defect area, and classifying the quality of the fiberboard. By analyzing the glossiness of the surface image of the fiberboard, the defect that the spatial domain feature is not obvious can be detected, so that the detection result is more accurate, and the product classification accuracy and the repair timeliness are improved; and the production efficiency is effectively improved.

The invention relates to textile equipment, particularly relates to a non-contact tension detection and feedback control mesh belt winding device, comprising a winding roller which drives the mesh belt to convolute to weave and is driven by an AC asynchronous motor; a charge couple device image sensor is arranged outside the mesh belt between two guide rollers, an ultrasonic wave sensor is arranged on the periphery of a ball of mesh belts, the charge couple device image sensor is connected with an image acquisition card, the ultrasonic wave sensor is connected with a signal acquisition card and the image acquisition card, the signal acquisition card is connected with a PC industrial personal computer connected with a motion control card, and the motion control card is connected with the AC asynchronous motor by a frequency converter; and a tension regulating device is arranged adjacently to the winding roller. The invention carries out non-contact detection on the variation of the tension force and the winding radius of the mesh belt and can not influence the tension force of the mesh belt due to the contact of the mesh belt, and the detection result is more accurate; and the tension force of the mesh belt is regulated by combining the operation of regulating the speed of the motor with the operation of driving a regulating roller by an air cylinder, thus the regulation of the tension force is faster and more effective.

The invention discloses a non-destructive testing method for optical glass polishing sub-surface damages. The method comprises the following steps of: preparing a non-destructive base at first, and then determining an incident angle for ellipsometry detection; calculating the optical constant of the non-destructive base according to ellipsometry parameters measured with the incident angle for ellipsometry detection, wherein the non-destructive base is defined as a substrate-environment structure, and a standard dielectric function is taken as a material physical model; defining a sub-surface damage surface of a sample as a multi-film optical model which comprises air, a coarse surface layer, a resedimentation layer and a substrate; building a multi-film material physical model through a mixed material effective medium model based on the multi-film optical model, and measuring the surface roughness and the depth distribution rule of surface sedimentation substances of the sample by virtue of the optical constant of the non-destructive base; and finally, carrying out inverse operation by a regression algorithm according to the ellipsometry parameters of the sample measured with the incident angle for ellipsometry detection, thus obtaining the depth of the sub-surface damage layer. The testing method is accurate, reliable, quick, efficient, flexible and controllable.

The invention discloses an image noisy point detection method based on a convolution neural network. The method includes the steps that sample images are collected, manual labeling and classification is conducted according to the types of noisy points, the sample images are input in the convolution neural network for training of classification models, and sample image blocks which are wrongly classified are collected in the classification process for secondary learning classification. The noisy points are labeled and classified in the manual and machine combination mode, supervised learning is achieved, learning accuracy of the convolution neural network is improved, and therefore the noisy points can be directly classified through the best trained classification model in the image noisy point detection process, and detection results are more accurate.

The invention discloses a test method for a storage, which mainly includes three test stages which are respectively read-write test, data wire test and address wire test and search for errors existed in the storage through step-by-step exclusive method which makes test results more accurately. Besides, the invention only needs utilizing the initial address and the final address to comprehensively detect whether the data wire has problems; and for detecting the address wire, the comprehensive test can be performed only by utilizing very few address locations. Moreover, through special arrangement of test data, test data groups and test address blocks, specific data wire and address wire on which an error is existed can be pointed out and the error can be reported in time, thereby greatly improving detection efficiency.

The invention discloses a train braking hardware-in-the-loop simulation testing stand and method, relating to the technical field of train brake tests. The testing stand comprises a testing-stand virtual part, a testing-stand hardware part and a data collection and signal conversion circuit, wherein the testing-stand virtual part comprises a man-machine interactive interface, a virtual train module for simulating a train and operational environment of the train, and a virtual electric braking module for providing a virtual electric braking force, the testing-stand hardware part comprises a driver controller for generating a train brake instruction and air brakes for providing air brake hardware required by tests, the number of the air brakes matches the marshalling number of a virtual train, and the data collection and signal conversion circuit is used for providing information exchange interfaces of the testing-stand virtual part and the testing-stand hardware part. The testing stand and method take full consideration of the complexity and high non-linearity of brakes, the testing result is accurate, and consumption of tests is greatly reduced.

The invention discloses a device for quickly detecting pesticide residues in fruits and vegetables, which comprises a color developing pool, a waste liquid pool, a sample liquid supply device, a color developing liquid supply device, a buffer liquid supply device, a substrate supply device, a sample liquid pump, a substrate pump, a color developing liquid pump, a buffer liquid pump, a valve system, a reactor and an electronic monitoring and control system. The device for quickly detecting pesticide residues in fruits and vegetables is used for quickly detecting the pesticide residues in fruits and vegetables, so that the interference of the color of a sample liquid of fruits and vegetables to be detected to the detection results can be fully eliminated. And meanwhile, the enzyme inhibition reaction and the enzyme-catalyzed substrate hydrolysis reaction are carried out in the same closed reactor, therefore, the impact of environment factor change is controlled to the minimum extent, and the detection results are relatively more accurate.

The invention discloses a method for quickly determining physical properties of flavoring essences for tobacco by the near infrared spectrum technology, which comprises the following technological steps in order: sample collection, namely more than 30 samples of the flavoring essences for tobacco are collected; content determination, namely the index contents of the collected samples is determined according to the method stated by the tobacco industry; spectrum scanning, namely near infrared transmitted spectrograms of the samples are collected by a transmitting and sampling module of a near infrared spectrometer; model formation, namely the near infrared spectrograms of the flavoring essences for tobacco are associated with corresponding indexes after pre-processing, the optimum spectral region range and the maximum principal divisor number are selected, and a mathematical model between the near infrared spectrums of the samples and the relative index contents is established and stored into a memory; and inspection of the flavoring essences for tobacco, namely the flavoring essences for tobacco of the same trademark undergo random sampling inspection, the near infrared transmitted spectrograms of the samples are acquired and compared with the mathematical model in the memory, and the physical indexes of the samples undergoing sampling inspection can be obtained. The method has the advantages of accurate testing result, simple and convenient operation, time and labor conservation, good repeatability of the testing result, high sensitivity and so on, and is suitable to be applied in the tobacco industry.

The invention provides a method for monitoring the expansion of fatigue crack in magnetic material by taking use of unusually changed magnetic signals, which pertains to the technical field of non-destructive detection. The method comprises the following procedures: a standard specimen is made with central crack of same material and heat treatment state as an object to be tested; after cycling preset times in fatigue testing, the central crack area of the standard specimen is scanned by a magnetic sensor with fixed liftoff, the normal component Hp(y) signal of the surface stray magnetic field of the scanned area is obtained; the length of the central crack 2a is measured; the max. value Delta Hp(y)max of the self unusually-changed magnetic signal generated by the fatigue crack is extracted as the detection threshold for the cycle numbers; the Delta Hp(y)max and 2a are linearly fit to get the Delta Hp(y)max-2a relationship formula; the object to be tested is inspected by the same liftoff with above magnetic sensor, the measured Delta Hp(y)smax is compared with the detection threshold; the expansion process of the fatigue crack is obtained according to Delta Hp(y)max-2a linear relationship. The invention is easy to operate, with accurate detection results, can monitor the expansion of crack dynamically under an unloaded state.

The invention discloses an adjustable-crack type temporary plugging acid fracturing steering performance testing device and an evaluation method, and belongs to the technical field of the petrochemical industry. The adjustable-crack type temporary plugging acid fracturing steering performance testing device includes a water storage tank, a plunger pump, an acid liquid tank, an adjustable simulatedcrack device, a filtrate metering electronic balance, and a computer control system which are sequentially connected. According to the adjustable-crack type temporary plugging acid fracturing steering performance testing device, a simulated crack is adjustable in size and form, and a parallel crack and a wedge-shaped crack of 0.5-8mm can be simulated; and both the simulated crack and temporary plugging acid liquid are heated to simulate the formation temperature, and the effect law of the crack size, shape, and the formation temperature on the temporary plugging effect can be studied. The effect of temporary plugging is mainly evaluated in performance by observing the change of upstream and downstream pressure difference, the filter loss, and the formation position of bridge blinding.

The invention relates to a content detection method for the Naoxintong capsule. The Naoxintong capsule comprises 66 parts of astragalus root, 27 parts of red peony root, 27 parts of root of red-rooted salvia, 27 parts of Chinese angelica, 27 parts of Szechuan lovage rhizome, 27 parts of peach kernel, 13 parts of safflower, 13 parts of prepared Frankincense, 13 parts of prepared myrrh, 20 parts of suberect spatholobus stem, 27 parts of twotooth achyranthes root, 20 parts of cassia twig, 27 parts of mulberry twig, 27 parts of earthworm, 13 parts of scorpion, and 27 parts of leech. The main chemical components, namely hydroxy safflower yellow A, paeoniflorin, ferulic acid, salvianolic acid B and ligustilide, in the Naoxintong capsule are simultaneously measured by the UPLC (ultra-performance liquid chromatography) technique. Single chromatographic analysis can be completed in 24 minutes. Chromatographic peaks of the main components are well separate, RSD (relative standard deviation) in both precision and receptiveness is smaller than 2.0%, and the quality of Naoxintong capsule can be controlled comprehensively.

The invention relates to a bionic mechanical abrasion resistance tester. According to the bionic mechanical abrasion resistance tester of the invention, a main swing component and an auxiliary swing component which are slidingly arranged on a main guide rod and an auxiliary guide rod respectively, and a driving device of the main swing component and the auxiliary swing component, and a control unit are utilized to drive a rigid supporting rod of which the lower end is provided with a shoe body fixing assembly to simulate the swing of a shank of a human being in walking; an abrasion disc assembly is arranged below the shoe body fixing assembly; a detected shoe body is fixed on the shoe body fixing assembly; and when the rigid supporting rod swings, the rigid supporting rod and the abrasion disc assembly rub each other relatively, so that the abrasion resistance of a shoe sole can be detected. Since the detection action of the bionic mechanical abrasion resistance tester is more approximate to the actual state of the human being in walking, detection results are more accurate.

The invention relates to a measuring method for cleaning function of clean and antibiotic ceramic in the field of measuring method of ceramic's function, which using the remained oil on the surface of product to measuring and less remained oily soil manifests easily to clean, vice versa, and which comprising the following steps: (1) cleaning the surface of product; (2) measuring product's quality B0 after dying; (3) dribbling oily soil into the surface of product and stewing; (4) rinsing oily soil by drinking water; (5) measuring product's quality B1 after dying again; (6) computing remained oily soil on the surface of product. A=104(B1-B0)/S, wherein, A for oily soil, S for product surface area, m2. The measuring method for cleaning function of clean and antibiotic ceramic is simple and useful. The invention is suitable for hydrophilic cleaning function test of daily-use ceramic ware and architecture healthy-use ceramic ware.

The invention discloses a transformer working state monitoring method. The method comprises the following steps: low-voltage user connecting transformers in different areas are subjected to area numbering, wherein a plurality of areas can exist; multiple low-voltage user connecting transformers in each area are then subjected to single item numbering; different types of transformers can be distinguished by adopting different number types; and the multiple low-voltage user connecting transformers in the same area are connected to a total signal line in the same area by sub-signal lines to transmit data back to a database. The monitoring method uses a vibration amplitude acquisition technology and a noise acquisition technology to detect the transformer, the fault diagnosis accuracy is improved, the fault type of the transformer can be analyzed timely, and according to the transformer monitoring method, a maintenance personnel can timely find out a transformer with potential safety hazards, monitoring and analysis on the transformer data in the system can be carried out in a long time, power supply safety is ensured, and power supply stability is improved.

A method for testing three-axis magnetic transducer includes generating zero magnetic field space, setting said transducer in said space, calculating out standard magnetic field values at three directions of tested three-axis magnetic transducer in sequence, utilizing obtained three field intensities to calculate standard value and making calibration on display value at three directions of three-axis magnetic transducer in sequence for obtaining test result.

The invention discloses a method for detecting phishing websites, which comprises the following steps: setting a website required to be protected; replacing corresponding characters in a domain name of the website with similar characters, wherein, a new similar domain name is formed through each replacement; carrying out matching on a domain name for accessing and a similar domain name by the gateway device so as to obtain a first similar value N of the two; comparing the first similar value N with a first preset value N1, if N is greater than N1, carrying out further direction, otherwise, ending the detection; respectively calculating the global features of an accessed page and the page of the website required to be protected, obtaining a second similar value N' through a similarity algorithm, comparing the second similar value N' with a second preset value N2, when N' is greater than N2, making an alarm; otherwise, ending the detection. The invention also discloses a system and a gateway device for detecting phishing websites. The method, system and gateway device for detecting phishing websites disclosed by the invention are accurate in detection, strong in pertinence, and good in protective effect.

The invention relates to an automatic voice summation tone detection method based on a deep neural network. The method comprises a step of training a deep neural network model for summation tone detection, wherein an input layer of the deep neural network model is characteristic information of voice and an output layer is probability output values for three states which are summation tone voice, voice of a single person and non-speech sound. The depth neural network model is used to perform summation tone detection on the automatic voice.

The invention relates to an infrared differential spectroscopy-based biogas detection and purification control system. The system has the advantages of high detection accuracy, low cost and long service life. The infrared differential spectroscopy-based methane detection and purification control system comprises a control center and a processing unit, and is characterized in that: A, a biogas tank is sequentially communicated with a dehydrating tower, a desulfurizing tower, a decarburization tower and a biogas storage tank through a pipeline; B, the pipeline inlets of the dehydrating tower, the desulfurizing tower, the decarburization tower and the biogas storage tank are provided with a sampling port respectively, and each sampling port is sequentially communicated with a pretreatment device and a gas chamber of an infrared detection center through a gas pipe; and C, a full-wave band infrared light source emits infrared light for detecting the gas to the gas chamber, an infrared detector receives and detects the infrared light after detection, and the acquired information is transmitted to the processing unit for data processing and is transmitted to the control center.

The invention discloses a gauge for measuring the size of a transformation position of a minor-diameter deep hole. The gauge comprises a casing and a measuring rod. A sidewall and an upper end face of the measuring rod are respectively provided with a guide head and a conical head, the casing is provided with a measuring passage, and the measuring rod and the conical head can move along the measuring passage. The distance from a lower end face of the guide head to a lower end face of the conical head is identical to the length of the measuring passage. A lower end of the casing is a conical face which is matched with a first conical guide section of the minor-diameter deep hole in shape, and the conical head is matched with a second conical guide section of the minor-diameter deep hole in shape. The bottom of the conical face, the top of the conical head and the transformation position of the minor-diameter deep hole are of a same diameter. A guiding slot is arranged on an upper end face of the casing, and the depth of the guide slot is larger or equal to a qualified inserting depth of the measuring rod. The guide head can move along the guide slot when rotating to the guide slot. The invention further discloses a method for the above gauge. By the aid of the gauge and the method, the size of the transformation position of the minor-diameter deep hole can be measured conveniently, quickly and accurately.

The invention discloses a method for measuring material nonlinearity based on phase object single-pulse reflection. Incident laser is divided into two beams through a first beam splitter: one beam of light is detecting light recorded by a first detector; the other beam of light passes through the phase object and then is focused on sample to be measured by lens; the intense pulsed light reflected from the surface of the sample to be measured is divided into two beams by a second beam splitter: one beam of light is directly recorded by a second detector; while the other beam of light enters a third detector through a small-holed aperture slot of which the center is coincident with the optical axis. The invention discloses a new method for measuring the optical nonlinearity of medium surface, which realizes the measurement of coefficient of nonlinear refraction and absorption, and can be widely applied in the research field of media interface nonlinear optical measurement and thin-film nonlinear optical measurement.

The invention provides a remote sensing image airplane target detection method based on a bounding box correction algorithm, and solves the problems that a traditional non-maximum suppression algorithm directly selects a bounding box with the highest score, and the detection bounding box is inaccurate, and a segmented airplane is marked by multiple boxes during detection, so that false alarm is caused. The method comprises the following steps: generating a training and testing sample set; preprocessing a training sample; overlapping and slicing the test samples to obtain preprocessed test samples; training an aircraft target detection model; detecting an aircraft target in the slice; carrying out bounding box correction on a detection result in the test sample by using a bounding box correction method; and generating an airplane target detection result graph. According to the invention, overlapping correction is carried out on the intersecting boundary frames on the same target by using an overlapping correction method, so that a more accurate detection frame is obtained. Fusion improvement is performed on the parallel adjacent boundary frames on the same target by using an adjacent fusion method, and multi-frame false alarms on the same target are removed in a detection result. The method is used for remote sensing image airplane target detection.

A test bar is prepared from following steps: preparing gold film electrode on insulating material by using micro electric mechanical system (MEMS); after electrode is treated through surface activation technique, a layer of nano metal grain is assembled on reaction zone on surface of the electrode through method of electrochemical deposition; then, reaction reagent is fixed on the said layer of nano metal grain so as to form reagent layer. When blood cholesterol of object to be analyzed through capillary action adds in drop wise to reaction zone of the test bar, detection system of electrochemical biosensor detects concentration of total cholesterol rapidly.

The invention relates to a crystal-free USB device clock calibration method and a calibration circuit, used for implementing the following steps: identifying and sampling an SOF packet on a USB bus, verifying the sampled SOF packet data, and outputting verification signals; counting the verification signals, comparing with a preset count value according to the counting result, and configuring a PLL parameter according to the comparison result; and clock-calibrating the RC oscillator of the USB internal clock and the PLL phase locked loop according to the PLL parameter. The invention has the beneficial effects that the information of the remainder part is retained, and the multi-phase sampling of the USB serial differential signal is carried out by using the fractional frequency dividing idea, so that the level characteristic of the USB serial differential signal can be accurately determined; all the information of the SOF packet is verified to ensure the reliability of SOF detection; the detection result will be more accurate; and in particular, the accuracy of the SOF packet detection of the invention is not affected by the internal clock bias to be calibrated.

The invention discloses a dual-reagent glycosylated hemoglobin detection kit. The kit comprises a reagent 2. The reagent 2 is prepared by labeling a monoclonal antibody (HbA1C monoclonal antibody), which can specifically recognize glycosylated hemoglobin, on latex microspheres through a chemical coupling method. The reagent 2 can replace the conventional reagent 2 (HbA1C monoclonal antibody) and the conventional reagent 3 (corresponding second antibody). The kit can be used to detect glycosylated hemoglobin, before detection, the kit does not to be shaken by an experimenter, the artificial operation error can be avoided, the kit is efficient and stable, and the detection results are more accurate. Furthermore, the reagent stability is high, so the application, storage and transportation become easier. A matured chemical coupling-latex labeling method is adopted, the kit can be produced massively, the difference between different batches of reagents can be controlled, and the cost is reduced. The detection cost of the method that uses the kit is lower, compared with that of HPLC method. The blank latex solutions produced by the mainstream factories can all be used to produce the kit. The kit can be promoted and used, and the economic burden for patients can be reduced.

The invention discloses a video contrast abnormity detection method and device. The video contrast abnormity detection method includes the following steps that: a video image to be detected is acquired from a video to be detected; brightness feature information of the video image to be detected is extracted through adopting a preset extraction mode; the brightness feature information is matched with a first kind of classification feature information, so that a first matching result can be obtained, wherein the first kind of classification feature information is determined by a plurality of pieces of brightness feature information which are correspondingly obtained in advance through extracting a plurality of video images of which the contract is abnormally high by adopting the preset extraction mode; and whether the first matching result satisfies a first preset matching condition is judged, so that whether the contrast of the video image to be detected is abnormally high can be determined. Compared with the prior art, the method and device of the invention can improve the accuracy of detection in which whether video contrast is abnormally high is detected.

The invention discloses a test device and method for the vertical initial motional impedance of a bucket foundation. The test device comprises the bucket foundation, a fan tower, a mass block, an extension thrust rod, a low-frequency electromagnetic exciter, an exciter fixing rack, a force sensor, a flange connecting disc, a soil pressure sensor, pore water pressure sensors, acceleration sensors, strain pieces, an energy-dissipation material, water drain pipes, a test iron box, steel angle ribs, steel plate reinforcing ribs, a loading frame, a power amplifier, a swept signal generator, a dynamic acquisition instrument and a computer. The test device is composed of a load loading part and a bucket foundation part, so that the vertical motional rigidity and the vertical motional impedance are measured. The low-frequency exciter is utilized for loading, and the motional rigidity of the bucket foundation is obtained by measuring the action force and displacement of a soil body. The motional impedance of the bucket foundation is obtained by applying vertical initial displacement to the mass block and measuring an attenuation curve. By adoption of the test device and method for the vertical initial motional impedance of the bucket foundation, the vertical initial motional impedance and influence factors of the bucket foundation can be accurately studied, and a basis is provided for further studying.

The invention relates to an aging testing method for simulating the practical operation environment of a rubber sealing ring. The method comprises the steps of applying stress to the rubber sealing ring, applying salt spray to the rubber sealing ring, applying high temperature to the rubber sealing ring, and applying low temperature to the rubber sealing ring. The method has the benefits that the salt-spray corrosion, stress effect and high and low temperature environment of the rubber sealing ring in the practical environment are sufficiently taken into account, the test process is more closer to the practical operation environment of the rubber sealing ring, and the test result is more accurate.

Being applicable to the data processing field, the invention provides a heat induction defect detection method and system. The method includes: extracting the image of a to-be-detected apparatus from infrared images, and identifying the equipment type and heat induction defect type of the to-be-detected apparatus; according to the equipment type and the heat induction defect type, selecting a corresponding heat induction defect analysis method for the to-be-detected apparatus, and selecting a heating analysis area from a heating part in the to-be-detected apparatus, wherein the heating analysis area includes a frame selected area and a line segment area; according to the heating analysis area and the heat induction defect analysis method, analyzing whether a heat induction defect appears in the to-be-detected apparatus. According to the invention, the specific heating analysis area and heat induction defect analysis method are determined according to the equipment type and the corresponding heat induction defect type of the to-be-detected apparatus, different analysis for different apparatuses is fully achieved, so that the judgment about the existence of a heat induction defect in the apparatus can be more accurate and reliable, and the heat induction defect detection result of the apparatus can be more accurate.

The invention discloses an image-sensor-based detection system of colloidal gold immunoassay and belongs to the field of the colloidal gold immunoassay. The detection system is as follows: a test strip placing platform for placing a test strip is arranged on a test strip detection platform; an illumination module and an imaging module are arranged above the test strip placing platform of the test strip detection platform; the illumination module consists of an illumination light source and a light homogenization unit arranged below the illumination light source; the illumination light source is a hollow annular light-emitting diode array light source formed by means that a plurality of light-emitting diodes are uniformly formed in an annular array; an image sensor of the imaging module uses a CMOS image sensor and the signal output end of the imaging module is electrically connected with the input end of a data collecting and processing device. The system adopts the annular matrix illumination and is provided with the light homogenization unit so that output light is homogenized and the detection result is accurate; the imaging module adopts the CMOS image sensor, so that the collection speed is high, and the requirement of dynamically tracking and observing the reaction process can be met.

The invention provides an LED flip-chip testing machine and a testing method, which can test an LED flip-chip timely at the chip stage of flipping the LED chip, wherein the provided testing data can be taken as a basis for light splitting and color separation, and can objectively reflect the LED chip flipping technology. The LED flip-chip testing machine comprises a support part, wherein a working plate is connected to an X-Y-Z triaxial driving device through the support part, and a gap capable of accommodating a plate-shaped solar panel is reserved between the working plate and the X-Y-Z triaxial driving device; and the working plate is transparent, and during test, the plate-shaped solar panel can move to the position under the working plate and collects light emitted by the LED flip-chip. The LED flip-chip testing method is characterized in that the working plate is transparent, and the plate-shaped solar panel is used for collecting light emitted by the chip under the working plate. The solar panel is small in size and adopts a plane structure, so that mounting and moving operation is convenient, and during test, positioning is more accurate, and light leak is not easy.