175results about How to "Overcoming sensitivity" patented technology

The invention relates to an antibody chip kit for diagnosis of various tumors. The kit includes an antibody chip, a tumor marker standard substance mixture, a biotin-labeled tumor marker detection antibody mixture, and a fluorescein Cy3-labeled streptavidin, wherein the antibody chip includes a substrate, 16 kinds of tumor marker specific antibodies fixed on the surface of the substrate, and two kinds of positive controls, and the tumor marker standard substance mixture is a freeze-dried mixture obtained by mixing 16 kinds of standard tumor marker standard substances together according to a certain amount. The kit can detect 16 clinical commonly used tumor markers, overcomes the defects of complicated operation, single detection index, low sensitivity and the like in the prior art, has the advantages of being cheap, convenient, sensitive, accurate, high in throughput, and less in amount of samples, and can be popularized and large-scaled in common laboratories.

The invention discloses a gallium nitride sensor, a preparation method and a multi-sensor system. The sensor comprises a heterostructure semiconductor substrate and a transistor which is located on the heterostructure semiconductor substrate, based on III-group nitrides and of an HEMT structure. Metal of the source electrode and metal of the drain electrode are both arranged on the portion, on the top layer of the transistor, of the semiconductor. A functionalized membrane obtained through functionalized processing is arranged on the surface of the grid electrode, and an induction area is formed in the exposed grid electrode area between the source electrode and the drain electrode. By detecting the current changes of the induction area, the concentration of to-be-detected substances making contact with the induction area is detected. By means of the scheme, the defects that in the prior art, sensitivity is low, the detection range is small and portability is poor can be overcome, and the advantages of being high in sensitivity, large in detection range and good in portability are achieved.

The invention discloses a lignosulfonate-polycarboxylic acid copolymerized composite high-performance water reducer and a preparation method thereof. The high-performance water reducer comprises the components in parts by weight: 5-30 parts of lignosulfonate, 20-30 parts of methyl allyl polyethenoxy ether, 2-5 parts of acrylic acid, 1-5 parts of ammonium persulfate, 4-15 parts of sodium hydroxide and water. The preparation method comprise the steps as follows: lignosulfonate, methyl allyl polyethenoxy ether and acrylic acid are directly synthesized under the action of ammonium persulfate, and the mixture is neutralized by sodium hydroxide to obtain the high-performance water reducer. As the high-performance water reducer is made from natural recycled materials, the high-performance water reducer is low in price, stable in performance, high in water-reducing rate, good in collapse prevention property, remarkable in improved effect, and is unlikely to bleed, separate and deposit after concrete is added, the quality problems of sand leakage, honeycomb-like phenomenon and leakage due to cracking in the middle and low-strength concretes are effectively avoided, and the safety performance is improved.

The invention discloses a preparation method of a MoS2 composite thin film with high hardness and loss abrasion on a surface of a substrate. The method comprises the following step: sequentially depositing a metal Ti priming coat, a TiN transitional layer and a Ti / MoS2 composite thin film on the surface of the substrate by using a high power pulse magnetron sputtering technology combined with optimized process conditions to obtain the MoS2 composite thin film which is strong in film binding force and high hardness and has low frictional coefficient and wear rate in various frictional environments. The nano hardness of the composite thin film is over 11GPa, the critical load value is over 60N, and the frictional coefficients in room-temperature air with relative humidity of 30%, 50% and 70%, an N2 atmosphere and a hydraulic oil environment are below 0.055. Therefore, the substrate is effectively protected in frictional reduction, and the composite thin film has a good application prospect.

The invention discloses a method for automatically recognizing face expressions based on multi-characteristic fusion, relates to a method for recognizing images. The method herein fuses images of face expressions, Gabor characteristics of images of major areas of face expressions and a multi-dimension ACILBP characteristic histogram. The method includes the following steps: pre-processing the images of the face expressions and the images of the major areas of the face expressions; separately extracting Gabor characteristics of the images of the face expressions and the images of the major areas of the face expressions, and imparting different weights, conducting fusion to obtain Gabor characteristics of two layers of images of the face expressions; using the ACILBP operator to extract the ACILBP characteristic histogram; conducting fusion to obtain the characteristic data of the face expressions; and using a SVM classifier to train and predict face expressions, and achieving automatic recognition of the face expressions. According to the invention, the method overcomes poor robustness under light and noise of prior art and the deficiency of low recognition rate due to insufficient utilization of partial information and whole information.

The invention relates to the technical field of information, in particular to an illumination lamp based on a microwave radar gesture recognition technology and a control method of the illumination lamp. The illumination lamp comprises a signal transceiving module, a gesture recognition module and a control module, and the signal transceiving module is used for periodically transmitting a microwave radar signal to a preset range and receiving a radar echo signal reflected by the hand; the gesture recognition module is connected with the signal transceiving module and is used for recognizing agesture motion and the gesture motion change of the hand according to the radar echo signal and generating a recognition signal corresponding to the gesture motion and the gesture motion change; the control module is connected with the gesture recognition module and is used for forming a control signal which corresponds to the recognition signal and is used for controlling illumination current according to the recognition signal generated by the gesture recognition module to adjust on and off and the brightness state of the illumination lamp. The illumination lamp is high in sensitivity, convenient to control and low in cost.

The invention provides a channel estimation method for a PDM-CO-OFDM (Pulse Duration Modulation-Coherent Optical-Orthogonal Frequency Division Multiplexing) system and relates to a system applying a multi-frequency code in a transmission process of digital information. The channel estimation method comprises the following steps of: designing and selecting a discrete guide frequency format at a sending end of the PDM-CO-OFDM system; sending information data and guide frequency data; utilizing a least square algorithm, namely an LS (Least Algorithm) algorithm, to carry out guide frequency point channel estimation on an orthogonal frequency division multiplexing mark by the information data and the guide frequency data received by a receiving end; carrying out linear interpolation on a frequency region and carrying out multi-wavelet de-noising on an obtained interpolation value; carrying out the linear interpolation on a time region; carrying out mark inner frequency averaging on an interpolation value result to finish combined de-noising; carrying out matrix inversion algorithm on a channel property subjected to the de-noising treatment; and finally, estimating sending data. According to the channel estimation method, based on a discrete guide frequency, the average combined de-noising of a multi-wavelet and a mark inner frequency region is used for carrying out channel estimation, so that defects that the calculation amount of the relative conventional OFDM channel estimation method is increased and the system expenditure is increased are overcome.

The invention discloses an intersection vehicle queuing length detection method based on an aerial video. The intersection vehicle queuing length detection method comprises the following steps: step 1, registering the aerial video; step 2, setting a vehicle queuing length detection region at an intersection entrance path; step 3, detecting a red light state; step 4, when a lane is at the red light state, detecting moving vehicles in the pre-set vehicle queuing length detection region by using a three-frame difference method; step 5, determining a vehicle queuing region; step 6, detecting the quantity of static vehicles of the vehicle queuing region and calculating an actual vehicle queuing length by using an actual distance Scale represented by a calibrated image unit pixel; and step 7, repeating the steps to realize continuous vehicle queuing length detection. According to the intersection vehicle queuing length detection method, the vehicle queuing length detection is carried out by the aerial video and mutual shielding and size changes of the vehicles in the monitored video are overcome, so that an intersection vehicle queuing condition is directly represented and the detection precision is extremely improved.

The invention provides a video semantic analysis method based on a self-adaption probability hypergraph and incremental semi-supervised learning. The video semantic analysis method based on the self-adaption probability hypergraph and the semi-supervised learning comprises the steps that (S1) a hypergraph model is established by means of a self-adaption probability hypergraph establishment method, (S2) the semi-supervised learning is conducted on the hypergraph model by means of the spectrogram segmenting principle, (S3) a semi-supervised model based on the self-adaption probability hypergraph is perfected by means of an increment mechanism, and (S4) semantic analysis is conducted on a tested video by means of the perfected hypergraph model. According to the video semantic analysis method based on the self-adaption probability hypergraph and the semi-supervised learning, the establishment of the self-adaption probability hypergraph and an incremental semi-supervised learning method are combined for use, the sensibility to a radium parameter when an ordinary hypergraph model is established is eliminated, and the accuracy and the robustness of the model are improved; in addition, under the incremental semi-supervised learning mechanism, semantic searching accuracy and semantic searching completeness are improved remarkably.

The invention relates to low temperature curable bottom filling adhesive with heat transfer function, comprising component A and component B; the component A is made from 100 parts of epoxy resin, 10-35 parts of modified epoxy resin, 0.5-5 parts of epoxy modified silicone oil, 5-15 parts of active diluent, 250-600 parts of heat transfer filler, 0.5-2 parts of a dispersant, 0.5-2 parts of a defoaming agent, and 0.2-4 parts of a rheological agent; the component B is made from 100 parts of liquid epoxy curing agent, 8-20 parts of a toughening agent, 250-600 parts of heat transfer filler, 0.5-5 parts of epoxy modified silicone oil, 0.5-2 parts of an adhesion promoter, 0.5-2 parts of a dispersant, 1-4 parts of a rheological agent, and 0.5-2 parts of a defoaming agent. The defects of low heat transfer coefficient, low temperature storage, high temperature curing and the like in existing filling adhesive can be overcome, and a chip packaged and cured by 3D MCP has high adhesive strength, good resistance to mechanical impact and good radiating performance.

The invention discloses a vibration sensing device based on a photoelectric oscillator, and the device comprises a laser, a one-in-two optical coupler, an optical phase shifter, a sensing head, a polarization controller, a Mach-Zehnder modulator, a two-in-one optical coupler, an energy-storage optical fiber, a photoelectric detector, a band-pass filter, a radio frequency amplifier, a 3dB power divider, a reference oscillator, a frequency mixer, a low-pass filter, an analog-to-digital converter, and an upper computer. According to the invention, one arm of the Mach-Zehnder modulator serves as the sensing head, thereby improving the sensitivity of vibration signal detection, and improving the detection precision. Compared with other passive optical fiber sensors, the device enables vibration information to be coded on the frequency of the vibration signal not the intensity, so the device is not sensitive to the impact from the power fluctuation of a light source. Compared with an optical fiber based on an optical fiber laser, the device is higher in sensitivity, and the sensitivity of the device is not related with the oscillation frequency. The device can be implemented through a low-cost low-frequency device, and is high in practicality.

The invention discloses a target tracking method based on fuzzy learning. The method mainly comprises the steps that firstly, a sparse representation algorithm is used to track a target in the aspect of a tracker; secondly, in the aspect of a detector, cascade classification detectors (a variance classifier, a collection classifier and a nearest classifier) are used to detect the target position; and finally, in the aspect of learning update, a fuzzy learner is used to integrate the output results of the tracker and the detector, and the final target position is acquired according to the membership of four constraints of time continuity, the spatial uniqueness , the similarity and the target size consistency. According to the target tracking method based on fuzzy learning, the real-time performance is ensured, and at the same time the adaptability to the target illumination change is great; and the discriminant ability of a learner is improved; the tracking accuracy and robustness of the algorithm are improved; and the method has important theoretical and practical significances for the research and practical application development of target tracking.

The invention relates to a seedling growing method, in particular to a seedling growing method for Scotch pine in northern sandy area. The invention is characterized in that seed source is selected; the seeds are arranged in snow or warm water to carry out the bud forcing treatment till that the seeds bud; seedlings are sowed by a level seedbed; meanwhile, soil amendment is fertilized; furthermore, during the period of seedling management, the management is carried out according to different growing conditions of the young seedlings; during the initial growing period of the young seedling, the nursery bed is watered; the temperature of the nursery bed is kept within 8-25 DEG C and the humidity thereof achieves 50%-80% of the water content in the field; furthermore, bordeaux mixture with the mass concentration of 0.8%-1.0% is sprayed once every 5-7 days and then the clean water is sprayed to clean the seedling; during the high growing period of the young seedling, the fertilizer is fertilized again thrice; during the stem growing period of the seedlings, the water content in the field of the nursery bed is kept within 35%-60%; the phosphate fertilizer is fertilized on the leaf surfaces; meanwhile, the winter protection disposal is carried out, thus improving the quality of the survival rate of the seedlings, improving the survival rate of the forest by 33.1% and generating considerable economical and social benefits.

The invention provides a vehicle license plate detection and recognition method based on combination of MSER and SWT. TO reduce influence of a shooting angle and illumination on license plate recognition and improve the license plate recognition accuracy rate under a complicated background, the method first performs MSER extraction and Canny edge detection, and performs MSER screening after an intersection of MSER extraction and Canny edge detection is taken according to license plate character characteristics; then SWT based on morphological processing is performed in areas after screening, areas after screening are aggregated, and combined with geometrical characteristics of a license plate, fine positioning of the license plate is completed; finally, segmentation correction is performedon connected domains in a successfully positioned area, a framework is extracted and normalization is performed, and the framework after normalization is matched with a template after refining and normalization; and HU invariant moment and grid characteristics are utilized to recognize the Chinese character of the first character, and jump point scanning, statistics and encoding are adopted to recognize numbers and letters. The vehicle license plate detection and recognition method provided by the invention has relatively a high accuracy rate and robustness of license plate detection and recognition obtained under a complicated background at multiple angles and under variable illumination conditions.

The invention belongs to the field of agricultural remote sensing application, and provides a model and a method for estimating the yield and shoot dry weight of soybean based on a spectral index. Four soybean yield models based on a two-band sensitive vegetation index are selected by comparing and analyzing the single-band, two-band and three-band vegetation indexes and a known vegetation index as follows: an exponential model based on NDVI (normalized difference vegetation index) (938, 642) at an R2 stage (a full-bloom stage), an R4 stage (a full-pod stage) and an R5 stage (pod-filling initial stage) of the soybean, an exponential model based on the NDVI (938, 642) at the R4 and R5 stages, an exponential model y=2.57e7.88x based on the NDVI (938, 642) at the R5 stage, and a power function model y=6280.97*6.78, and verification on the models is carried out. In addition, a common core waveband of the yield and the shoot dry weight of the soybean is determined, and the predictability of the vegetation index based on the common core waveband for the shoot dry weight can be further confirmed. By adopting the method, the yield and the shoot dry weight of the soybean can be rapidly and nondestructively estimated. The method and the model are applicable to the large-scale breeding of the soybean and particularly applicable to the high-yield breeding of the soybean in ecological areas of the middle and lower reaches of Yangtze River.

The invention discloses a method for detecting a trace amount of oxytetracycline by utilizing a CdTe quantum dot fluorescent probe. According to the invention, CdTe quantum dots are taken as the fluorescent probe, and oxytetracycline and CdTe quantum dots are utilized to be combined to form a new composite system through electrostatic interaction to cause the quenching of fluorescent light of CdTe quantum dots, so that a method for detecting the trace amount of oxytetracycline is established. A favorable linear relation is formed by the quenching amount of fluorescent light of CdTe quantum dots and the concentration of gibberellin within the concentration range of 8*10<-9>-2*10<-7> mol / L, and the detection limit of the method is 2.72*10<-10> mol / L. According to the invention, defects of low sensitivity, long reaction time, poor stability and narrow application range during detecting in the prior art are overcome, the sensitivity and selectivity are better improved, and the detection to the low-concentration oxytetracycline is more convenient and faster.

The invention discloses a primary direction neural network system with a four-layer feedforward structure. The primary direction neural network system comprises an input layer, a first hidden layer, a second hidden layer and an output layer, wherein the input layer comprises D neurons, the first hidden layer comprises K groups of neurons, each group of the neurons comprises three neurons, the second hidden layer comprises K neurons, the output layer is a neuron, and both the D and the K are natural numbers; the input layer is used for receiving the D dimensional vector, and each neuron correspondingly receives one component in the D dimensional vector; the first hidden layer is used for mapping the D dimensional vector received from the input layer to the neurons in the second hidden layer, and each group of the neurons in the first hidden layer corresponds to one neuron in the second hidden layer; the second hidden layer is used for mapping the 3K dimensional vector received from the first hidden layer to the neurons in the output layer; and the output layer performs biased w0 linear weighting to the result of the second hidden layer and then outputs the result. The invention overcomes the extremely small difficulty of being trapped in local and the sensitivity to the noise.

Relating to laser Raman spectrum detection, the invention discloses a method for enhancing Raman spectroscopy by anion modified nanoparticles. According to the method, metal nanoparticles are prepared into a core, a full layer is strongly adsorbed with surface modified nanoparticles with anions as a modification layer; a nanoparticle sol and a to-be-measured biomolecule are mixed uniformly; and surface-enhanced Raman spectroscopy detection is conducted directly. The method for enhancing Raman spectroscopy by anion modified nanoparticles provided by the invention has the advantages of simple operation, fast speed, low cost, high sensitivity, good reproducibility, effective maintaining of the native conformation of biomolecules, and accurate and reliable result, thus being suitable for biomolecule study.

The invention provides an efficient anti-sticking UV release paint applied to a pressure-sensitive adhesive as well as preparation and application of the efficient anti-sticking UV release paint. The paint comprises the following ingredients in percentage by weight: 3-8% of a photoinitiator, 3-5% of an auxiliary initiator, 20-40% of an active diluent, 15-45% of organic silicon-acrylic resin and 25-45% of non-silicon acrylic resin. The system of free radicals and the cationic composite initiator is adopted, a sensitization effect realized through the synergistic effect of the free radicals and the cationic composite initiator better solves the problems of oxygen inhibition and slow curing caused by the single initiator, meanwhile, the photo-curing speed is high, the solubility is good, the absorption wavelength reaches 390 nm and the paint is an efficient, energy-saving and environment-friendly photo-cured release paint.

The invention discloses a femtosecond time resolved excited Raman spectrum system, which can be used for detecting the vibration dynamics of the molecular electron base state, can also be used for detecting the vibration dynamics of the molecular electron excitation state, and can give out rich information relevant to molecular electron excitation state properties, particularly vibration excitation state dynamics. The Raman spectrum system generates excitation light, Raman pumping light and detection light; a spectrum instrument can perform double chopping modulation on the excitation light and the Raman pumping light, and can control the time sequence of the excitation light, Raman pumping light and detection light reaching the samples. The wavelength of the Raman pumping light can be continuously tuned; the whole-frequency excited Raman spectrum from visible light to near ultraviolet waveband can be obtained; in addition, the proper wave length can be selected to effectively avoid the fluorescence interference; the intensity of the excited Raman spectrum can be improved under the resonance conditions; the defects of low sensitivity and fluorescence interference of the conventional Raman spectrum are overcome.

The invention particularly relates to a data clustering analysis method based on a genetic algorithm. The data clustering analysis method based on a genetic algorithm comprises the following steps: firstly, selecting an initial population from a sample set to be clustered; executing a genetic algorithm on the selected initial population; executing k-means operation on the new population generatedafter the genetic algorithm is executed; and repeating step (A) to step (C) until the optimal solution of the clustering problem is found. In the data clustering analysis method based on a genetic algorithm, local optimization of the K-means algorithm and global optimization of the genetic algorithm are combined, the optimal clustering number and the initial centroid set are finally obtained through multiple times of selection, intersection and variation genetic operation, the locality and sensitivity to the initial clustering center of a traditional K-means algorithm are overcome, and effective classification of data is achieved.

The invention discloses a molecular probe for detecting African swine fever virus, a kit and application thereof, so as to solve the technical problems of poor reliability and poor practicability of the existing detection technology. The nucleotide sequence of the molecular probe for detecting African swine fever virus is shown as in SEQ ID NO.2. An RT-PCR reaction solution for detecting African swine fever virus is prepared, lyophilized powder for the RT-PCR reaction solution for detecting African swine fever virus is prepared, an African swine fever virus detection kit is prepared, and an African swine fever virus qualitative and quantitative detection method is provided. The designed molecular probe and the built RT-PCR detection method have the characteristics of high sensitivity, goodrepeatability and strong specificity, the reliability and the usability of ASFV detection by the PCR technology are improved, pollution-free and high-flux detection is realized, and ASFV nucleic acids in susceptible animals and related animal products can be detected.

The method includes steps: first, preparing histogram statistics for mismatched pixels of moving forwards compensation block in current input image; next, calculating statistic of heterotactic characters of histogram for mismatched pixels of moving forwards compensation block between adjacent frames; then, carrying out determining procedure for switching scenes based on the said statistic, and outputting the determined result. Based on difference detection technique of moving compensation, the invention overcomes disadvantages: sensitivity to noise and movement caused by the difference algorithm for pixels located at same positions in adjacent frames; and negligence of space information of video object occurred in conventional decision method of analyzing statistical histogram of pixel intensities. Advantages are: lowering probability of false alarm and probability failed to report in detection of switching video scenes as well as raising precision and correctness of determination.

The invention relates to a power-system direct current power transmission line relay protection field and especially relates to a single-terminal electric-quantity whole-line quick action protection method which can be used to identify faults in and out of an HVDC transmission line area. An amplitude of a specific frequency component in a single-terminal direct-current filter branch is used to realize discrimination of the faults in and out of the area. The method of the invention is used to protect a single-terminal whole-line quick action in the direct current power transmission line of the power system. The protection can be achieved only by using the single-terminal electric quantity. The method has a low requirement to a sampling frequency. An algorithm is simple. Sensitivity is high. The method can be realized during engineering. The action speed is fast, selectivity is good and reliability is high. A theory is completed and is easy to be adjusted. The protection method of the invention can replace current traveling-wave protection which is taken as the main protection of the direct current power transmission line. And the method is especially suitable for realizing the whole-line quick action protection of the ultra / super HVDC power transmission line by using the single-terminal electric quantity.

The invention relates to an ultrasonic-acoustic emission detection method for a composite material. The method can detect internal defects of the composite material according to an acoustic emission signal excited by pulse ultrasonic waves in the composite material; an emission transducer and a receiving transducer adopt an asymmetric mode and two layouts such as a separation type and an integration type and are arranged on the same side or any side of a part of the composite material; the positions and the attitudes of the emission transducer and the receiving transducer are selected within 0-360 degrees, so that multi-channel ultrasonic-acoustic emission detection can be realized; defects can be judged and positioned according to a time domain feature of an acoustic emission signal in the detected composite material. According to the method, the composite material detection capability of ultrasonic-acoustic emission is greatly improved; the detection thickness is 50mm, and the detection sensitivity to the defects which have the diameters being phi3mm and are 0.13mm and 49mm deep is achieved; the ultrasonic-acoustic emission detection resolution, a surface detection blind area and the detection capacity are obviously improved.

The invention relates to a device and method for detecting the weak absorption of an optical element. The device comprises a pump laser, a nonlinear crystal, a dichroscope, a reflecting structure, a first focusing lens, a second focusing lens, a filtering element, a pinhole and a photoelectric detector. The nonlinear crystal is arranged at the laser emitting end of the pump laser, and an included angle exists between the dichroscope and the nonlinear crystal. The reflecting structure is used for reflecting a frequency doubling laser emitted by the dichroscope to the surface of the optical element. The second focusing lens is used for focusing the frequency doubling laser emitted by the reflecting structure to the surface of the optical element. The filtering element and the pinhole are sequentially arranged between the optical element and the photoelectric detector. The device and method for detecting the weak absorption of the optical element can overcome the defect that due to the different phases of pump lasers and detection lasers, direct-current components affect test results in a traditional mode, and the device and method for detecting the weak absorption of the optical element are particularly suitable for measuring short-pulse pump lasers and greatly improve the sensitivity and the precision of weak absorption tests of the optical element.

The invention discloses a meteor detection method based on an incoherent scattering radar, and the method comprises the following steps: (1), dividing original baseband orthogonal two-path sampling data into two parts, enabling one part of the data to serve as the input data of meteor head event detection, enabling the other part of the data to serve as the input data of meteor trail event detection, and enabling the two parts of event detection to operate in parallel as independent steps; (2) carrying out meteor event detection; and (3) carrying out meteor parameter extraction. According to the meteor detection method based on the incoherent scattering radar, the sensitivity to a radar beam arrival angle is overcome, the coverage range of a detectable meteor area is widened, meteor eventscan be detected in real time, and meteor parameters such as the height, the radial speed, the radial deceleration and the equivalent scattering sectional area are extracted.

The invention discloses an MIMO radar multi-measurement vector DOA estimation method based on iterative near-end projection. The method comprises the following steps: converting high-dimensional echodata into a low-dimensional space to reduce the spatial dimension, performing singular value decomposition on the data subjected to dimension reduction, extracting a signal subspace to reduce the timedomain dimension, representing the non-convex non-smooth sparse optimization problem in the MIMO radar multi-measurement vector DOA estimation by using a near-end function optimization model, and then obtaining a near-end operator through an extrapolation step and an SCAD function in the iterative process to solve the optimization problem. According to the invention, the DOA estimation performance of the coherent information source under the conditions of low snapshot and low signal-to-noise ratio is superior to that of the existing algorithm.

The invention discloses an arc heater water leakage fault diagnosis system based on atomic emission spectroscopy. The air is heated by an arc heater and expanded and accelerated by an effuser to formhigh temperature air flow. The flow field spectrum formed by the arc heater is focused by a spectrometric compression film. The focused flow field spectrum is transmitted to a spectral measurement module through an optical fiber. The spectral intensity of the oxygen atomic characteristic spectrum and the spectral intensity of the background spectrum are measured by using the spectral measurement module and then the relative intensity is calculated. The relative intensity is compared with the intensity threshold by a data processing terminal, and the water leakage fault of the arc heater is judged according to the comparison result. When the judgment result indicates water leakage, power supply to the arc heater is disconnected. Rapid diagnosis of the low enthalpy arc heater water leakage fault can be realized so that safe operation of the heater can be guaranteed; and the operation state of the heater can also be monitored in real time so as to provide reference to analysis of the long-term operation stability.