610results about How to "Improve analytical accuracy" patented technology

The invention relates to a slice image processing method and device, computer equipment and a storage medium. The method comprises the following steps: segmenting a slice image into a plurality of cell images according to acquired image segmentation parameters; inputting the cell image into the trained convolutional neural network model; obtaining classification probability data, and according tothe classification probability data, determining a classification category corresponding to the cell image, then counting the number of cell images of an abnormal category in the slice image, calculating the proportion data of the counted number of cell images in the total number of cell images, and when the proportion data is greater than a preset proportion threshold, marking the slice image asan abnormal slice image. On one hand, detailed analysis is carried out by segmenting the slice image. The overall analysis accuracy of the slice image is improved. On the other hand, classification accuracy of a single cell image is improved by adopting an artificial intelligence convolutional neural network, so that the analysis accuracy of the slice image is improved.

A continuous casting powder X-ray fluorescence spectrometry mainly comprises the following steps of producing calibration samples by mixing a standard blast furnace slag sample, a standard fluorite sample and a standard coke component sample together, producing the synthetic standard blast furnace slag sample, the synthetic standard fluorite sample and the synthetic standard coke component sample into standard sample photos which can be analyzed by an X-ray fluorescence spectrometer in a melting method, scanning one or more standard sample photos which are melted well with the X-ray fluorescence spectrometer, measuring all the standard sample photos with the X-ray fluorescence spectrometer according to the fundamental analysis condition, producing samples to be measured into sample photos which can be analyzed by the X-ray fluorescence spectrometer in a melting method, analyzing the samples to be measured and conducting X-ray fluorescence spectroscopy analysis to the samples to be measured with the X-ray fluorescence spectrometer.

The invention discloses a breast cancer cell characteristic analysis system based on deep learning. Based on the deep learning, the system constructs a multilevel convolutional neural network to extract multilevel features, so higher analysis accuracy can be realized; an activation function of a model in the breast cancer cell characteristic analysis system disclosed by the invention is an unsaturated ReLU function, which has faster convergence properties; a pooling layer in the breast cancer cell characteristic analysis system disclosed by the invention adopts an overlapped pooling operation, it can be proved by cross validation that, compared with a traditional non-overlapped pooling layer, the overlapped pooling can further improve the analysis accuracy; and the breast cancer cell characteristic analysis system disclosed by the invention adopts a training mode of sparse autocoder pre-training and Dropout fine tuning to effectively reduce the over fitting of the model and enhance the generalization ability of the trained model, so that the analysis accuracy can be further improved.

The invention discloses an X-ray fluorescence spectroscopy analysis fusion sample preparation method of an aluminum, manganese, calcium and iron alloy, which comprises the steps of: firstly, coating a uniform protection wall layer in a platinum crucible by using lithium tetraborate, then placing an aluminum, manganese, calcium and iron sample, an oxidizing agent, a lithium borate solvent and a releasing agent into the platinum crucible in which the protection wall is coated, uniformly mixing, and then pre-oxidizing the aluminum, manganese, calcium and iron sample; and finally, melting the pre-oxidized aluminum, manganese, calcium and iron sample, and cooling to obtain an aluminum, manganese, calcium and iron sample glass fuse piece, wherein the aluminum, manganese, calcium and iron sample is a detection sample or calibration sample of the aluminum, manganese, calcium and iron alloy, and the sample glass fuse piece is a detection sample glass fuse piece or calibration sample glass fuse piece of the aluminum, manganese, calcium and iron alloy. On the premise of no corrosion to the valuable platinum crucible, the glass fuse piece can be prepared, and the prepared sample is excellent and uniform, thus the mineral effect and the grain size effect are completely eliminated.

The embodiment of the invention provides a network live broadcast monitoring method and device and a data processing method. The monitoring method comprises the steps of determining whether obtained interaction information of a target live broadcast room meets a preset suspected risk condition or not; if yes, selecting at least one video frame from the live video of the target live broadcast roomas a suspected risk frame according to the interaction information; and performing live broadcast risk analysis on the target live broadcast room according to the suspected risk frame to obtain a riskmonitoring result. According to the invention, firstly, a suspected risk live video is identified based on interactive information, and then live broadcast risk analysis is performed on video framesrelated to the interactive information in the live broadcast video; by introducing interactive information of a target live broadcast room, due to the fact that the recognition efficiency of the interaction information is high, the live video with the suspected risk can be rapidly positioned, live risk analysis can be conducted on the live video in a targeted mode, the live risk analysis efficiency and accuracy can be improved, and the situation that risk live broadcast violation playing occurs due to video frame extraction omission is reduced.

Disclosed herein is an optical detection method and optical detection apparatus, the apparatus including: a light irradiation section configured to irradiate a laser beam upon one of fine particles which are successively fed in a flow path; and a light detection section configured to detect fluorescent light and / or scattered light generated from any of the fine particles upon which the laser beam is irradiated; the method including the steps of: irradiating a laser beam upon one of fine particles which are successively fed in a flow path; and detecting fluorescent light and / or scattered light generated from the fine particle; wherein the laser beam being formed as a pulse laser beam whose pulse intensity is modulated such that one laser beam or two or more laser beams having different wavelengths are irradiated by a plural number of times upon one fine particle with the intensity varied.

The invention discloses a method for quickly measuring the content of phosphorus, lead and zinc in industrial ferrous dust. The method comprises the steps of sample decomposition, setting the instrument analysis parameters of an ICP (Inductively Coupled Plasma) spectrometer, interference correction, making a working curve, and analyzing the content of to-be-measured elements in a measured sample by using an ICP spectral method, wherein the sample decomposition comprises sample pre-decarburizing treatment and sample pressure-microwave digestion. The invention provides a new method for improving the speed and the accuracy in a chemical test of hazardous elements in metallurgical dust.

The invention provides a tourism data analysis system and method, and the system comprises a data collection module which is used for collecting data of various scenic spots; the data storage module is used for classifying and storing data of various scenic spots; the data analysis module is used for establishing a tourist identity information base, a vehicle information base and a tourist consumption information base; generating tourist track data, scenic region heat distribution data, scenic region public praise indexes and tourist comment data; carrying out multi-source data fusion on the tourist identity information base, the tourist track data, the tourist comment data and the tourist consumption information base to obtain tourist portrait data; generating scenic area decision data byusing the scenic area public praise index, scenic area heat distribution data and tourist portrait data; and the external interface module is used for externally providing the acquired various typesof data in an interface access mode. By applying the embodiment of the invention, the efficiency and accuracy of tourism data analysis are improved.

The invention discloses an intelligent substation message simulation testing and association message analysis system and method. The system comprises an SCD (Security Coding Device) file analysis module, a configuration reading module, a template and case management module, an intelligent device interface simulation module, a message recording module and a user interface module, wherein the SCD file analysis module is used for analyzing the configuration descriptive SCD files of a substation; the configuration reading module is used for extracting basic information required by cases of relevant tests; the template and case management module is used for collecting and managing test cases and standard response templates aiming at the test cases; the intelligent device interface simulation module simulates and sends relevant messages accordant with the IEC61850 standard; the message recording module is used for capturing or responding to the sent messages and generated events and alarming; and the user interface module is used for collecting the test cases and the templates and displaying analysis results. According to the invention, artificial intelligence is introduced in rapid analysis on results through a template method, the message simulation capability of the system is used for flexibly designing test cases, intelligently recognizing and realizing test conditions, thus greatly improving the test and analysis efficiency and accuracy.

The invention discloses a method for graphically monitoring a running state of a transformer of a distribution network. The method comprises the following steps: displaying a real-time running state and data of a distribution transformer on a WEB window page in a graph manner after treating and calculating through a software system according to the running data of the monitored distribution transformer collected and sent back by an on-site measuring terminal device, wherein the graph shows the geographic information of the location of the distribution transformer and displays the current state of the distribution transformer with different colored blocks, and the colors of the colored blocks can directly reflect the high-low degree index of the load factor corresponding to the distribution transformer; and classifying and grading various distribution transformers according to the high-low degree index of the load factor and corresponding state color of each distribution transformer. By using the method, the running state of the transformer of the distribution network can be monitored in real time, so that the work loads of on-site inspection tour and long-time load measurement are greatly reduced, the hidden danger for the running of the distribution transformer is accurately and timely found out, a reforming plan can be made earlier, and the operation and maintenance levels of the device are improved.

The invention relates to a smarter-vision-based high-performance ore granularity visual inspection system. Detection steps are as follows: a high-speed and high-definition camera is installed above anore conveyor belt; an ordinary image and an image with structured light of broken ore on the ore conveyor belt can be taken by the camera; information of the ordinary image and the image with the structured light is transmitted to a system host for detection, ore particles are cut, ore granularity information is extracted for statistics and analysis, and the ore granularity information is outputted to a main controller of an ore crusher; the main controller adjusts various control parameters of the ore crusher according to the requirements of the production process, and the ore crusher operates according to new control parameters; after the ore crusher operates according to the new control parameters, the broken ore again appears in the image of the camera, and the closed-loop control ofthe ore crusher is realized. The smarter-vision-based high-performance ore granularity visual inspection system can effectively solve the problems that an ordinary camera can only collect two-dimensional image information, and an ore particle size analyzer cannot distinguish an damaged or water-stained empty conveyor belt from a conveyor belt scattered with the ore, so that the ore granularity information detection and analysis are inaccurate.

The invention discloses an online remote calibration method for an intelligent electric meter based on measurement data analysis. The method is based on power data. the method comprises the followingsteps of preprocessing an original measurement data time sequence by utilizing an improved fuzzy C-means clustering technology, removing measurement data under a light load working condition, and taking the measurement data as an input variable of an on-line remote calibration model; establishing an intelligent electric meter error solving equation on the basis of computing the line loss of a zonearea; estimating an intelligent electric meter operation error on the basis of a limit memory recursive least squares algorithm; checking and analyzing an intelligent electric meter error estimationresult; and performing a service assistant decision making through combination with an on-line remote calibration result of the intelligent electric meter. According to the method, the precision, efficiency and practicability of intelligent electric meter error estimation are improved based on power data analysis.

The invention relates to the technical field of computers, and discloses a method and a device for intelligently analyzing electronic day-to-day journals, as well as a processor. According to the method for intelligently analyzing the electronic day-to-day journals, a to-be-analyzed journal file is decomposed into individual electronic day-to-day journals item by item, and an abnormal electronic day-to-day journal is acquired; further, the electronic day-to-day journal with an incorrect journal sequence can be acquired according to a journal record template directed graph, and the electronic day-to-day journals are divided according to a business journal record directed graph and an executive business; and an abnormal operation and an normal operation are determined, and corresponding operating state information of self-service equipment is acquired. In the scheme, not only is the correctness of each electronic day-to-day journal analyzed, but also the logic sequence relationship of the electronic day-to-day journals is checked, the degree of automation is high, and the analysis accuracy of the electronic day-to-day journals is higher; and compared with the prior art, neither manpower nor material resources are consumed greatly, and the consuming time is short.

The invention relates to a network application protocol analysis method, in particular to a seven-layer application protocol depth analysis method based on a DPI. The method comprises the following steps of: obtaining various application software data packets by using a packet capturing tool; determining which kind or which one obtained application software belongs to through the following steps: (1) adding an application protocol analysis layer into a tcp / ip protocol stack, (2) establishing an application protocol matching model by using a depth packet analysis arithmetic, and (3) obtaining a network data packet and analyzing characteristic application of the data packet to obtain application protocol information of the network data packet; carrying out protocol matching through the application protocol matching model with a multi-mode matching arithmetic; and determining which kind or which one the protocol type belongs to. According to the application protocol analysis method based on the DPI disclosed by the invention, various application software data packets are obtained by using the packet capturing tool; the protocol matching is carried out by using a high-efficient DPI depth packet detection technology with the multi-mode matching arithmetic; the rapid application protocol analysis is carried out on the data packets; and the analysis efficiency and accuracy are increased.

The invention relates to a mobile terminal health information acquisition and health analytical method based on cloud computing. A health information acquisition instrument collects health information and transmits the health information to a mobile terminal, the mobile terminal uploads the health information to a cloud server through a wireless network, and a health analytical platform reads the health information received by the cloud server, obtains health analytical suggestions and then downloads to the mobile terminal corresponding to the health information. Various health information acquisition instruments are adopted at a home terminal of the mobile terminal to collect the health information which is transmitted to the cloud server through the mobile terminal to store, long-distance acquisition of the health information is achieved, and efficiency is improved compared with the existing manual on-site collection method. Simultaneously, various health information is uniformly uploaded to the cloud server to bring convenience to management of system data, long-distance concentration expert consultation and health suggestion analysis are achieved, the health suggestion analysis can be transmitted to the mobile terminal in a long-distance mode through the wireless network, and efficiency and accuracy of the expert consultation and the health suggestion analysis are improved.

The invention relates to buried pipeline protection technology, in particular to stray current detection and analysis technology of buried pipelines. The invention discloses a stray current tester and a stray current analysis method, which can reflect the distribution situation of stray current on the ground more comprehensively and intuitively, and provide a scientific evidence for the protection of the buried pipelines. The stray current tester comprises a data acquisition device and at least two detection channels, wherein each detection channel can measure a path of voltage; the data acquisition device is used for recording each path of voltage and drawing a graph according to recorded voltage data to display; and the voltage comprises a direct current component and an alternating current component. The stray current analysis method utilizes a fuzzy expert system of a computer to comprehensively analyze acquired various data, and an obtained result is more accurate and comprehensive.

The invention discloses a malicious software feature fusion analytical method and a system based on shared behavior segments. The method includes deploying, collecting and analyzing nodes of malicious software, and constructing a distributed hash table (DHT) module; collecting samples of the malicious software and segmented into segment sets, and calculating local statistical properties; sharing to the DHT, gathering global features of the behavior segments, and returning to source nodes; the source nodes calculating candidate neighbor node sets and performing similar calculation of behavior characteristics through remote nodes of the candidate neighbor node sets to construct an adjacency relation diagram of the behavior characteristics; and generating an aggregation three for aggregation based on the adjacency relation diagram of the behavior characteristics, and outputting root behavior characteristics. The system comprises a plurality of nodes, each node comprises a characteristic segmenting module, the DHT module, a behavior segment synergy sharing module, a neighbor behavior characteristic discovering module and a behavior characteristic gradual aggregation module. The method and the system have the advantages of being high in analytical accuracy and performances and good in expandability.

The invention discloses a method for analyzing influences of a background ionosphere on GEO SAR imaging. The method includes the steps that a GEO SAR echo signal model influenced by a time-varying ionosphere is established, image offset and defocusing phase which are used for analyzing influences on imaging are derived on the basis of the model, and then judgment of the influences on imaging is performed; during analysis, GEO SAR echo signals influenced by the time-varying ionosphere and the measuring data of the ionosphere total electron content TEC on the propagation paths of the GEO SAR signals are obtained, polynomial fitting is performed on the TEC measuring data three times, a constant term and coefficients of first three items are obtained, echo signal parameters and TEC coefficients are substituted into a derived expression of the image offset and the defocusing phase, and compared with a threshold value, the influences of the background ionosphere on GEO SAR imaging are obtained. According to the method, modeling is performed specifically and is accurate, and imaging analysis is performed on the basis of the accurate model so that the analysis accuracy can be improved.

The invention provides a multi-class target object detection method and device, an electronic device, a storage medium. The multi-class target object detection method comprises: training a first neural network model to enable the first neural network model to output a quasi-first target frame and a quasi-second target frame, the quasi-first target frame indicating a first class of target objects,and the quasi-second target frame indicating a second class of target objects; training a second neural network model to enable the second neural network model to output second-level confidence coefficients and attribute information of the quasi-first target frame and the quasi-second target frame; taking the trained first neural network model and the trained second neural network model as a multi-class target object detection model; and inputting the to-be-detected picture into the multiple types of target object detection models, and determining the superiority according to the second-levelconfidence coefficients and attribute information of the quasi-first target box and the quasi-second target box output by the multiple types of target object detection models. According to the methodand the equipment provided by the invention, simultaneous detection of multiple types of objects is realized.

The invention belongs to the technical field of analysis and in particular relates to a device and method for combining electrochemistry with electrospray ionization mass spectrometry. The device is composed of an electrochemical reactor and an electrospray ionization ion source. The electrochemical reactor comprises a work electrode chamber and an auxiliary electrode chamber which are isolated from each other through a diaphragm, the diaphragm has the effects of preventing or delaying a neutral substance from exchanging between the electrode chambers and allowing an ion to exchange with an electric ion. A solution sample can be subjected to electromechanical reaction under the effect of an electrochemical voltage loaded between the work electrode and the auxiliary electrode when flowing in the work electrode chamber. An electrochemically reacted product is protected by the diaphragm and then forms an air-phase ion under the effect of the electrospray ionization work voltage, and the air-phase ion is analyzed by a mass spectrometer. According to the invention, the electromechanical reaction efficiency is higher, the electromechanically reacted product is simple, and the response speed is high, thus the device and method provided by the invention adapt to the rapid, accurate and high-flexible advantages of the quantitative analysis and the mass spectrum analysis.

The invention discloses a method of detecting aging degree of an anti-pollution flashover coating of room temperature vulcanized silicone rubber. The method comprises the following steps of sampling to obtain a sample wafer; testing one surface, exposed to the air, of the sample wafer by adopting an attenuated total reflection fourier infrared spectrometry to obtain an infrared spectrogram; and judging the aging degree of the coating according to the ratio of a peak value of a 1008cm-1 absorption peak and a peak value of a 786cm-1 absorption peak and a peak value of a 1008cm-1 absorption peak through the infrared spectrogram. The detection method provided by the invention can carry out ration judgment on the anti-pollution flashover coating of the room temperature vulcanized silicone rubber through a few of sample wafers, the specific sampling is not needed, the method is simple and practical, the analytical precision and accuracy are high, and the reproducibility is good.

The invention discloses a wafer on-line detecting method, the steps of the method comprise providing a to-be-detected wafer, wherein a plurality of units are periodically arrayed on the to-be-detected wafer, and each unit is internally provided with a detecting area, determining whether the detecting areas are in the corresponding units or not, determining the distribution of the detecting areas on the to-be-detected wafer according to the position of the detecting areas in the corresponding units, selecting a plurality of to-be-detected units according to distribution of the detecting areas on the to-be-detected wafer, detecting the detecting areas of the pluralities of to-be-detected units, and obtaining detecting data. The invention further discloses a corresponding wafer on-line detecting device, after utilizing the wafer on-line detecting method and the on-line detecting device of the invention, obtained detecting data can relatively accurately stand for the detecting data of the whole wafer, thereby increasing the analyzing accuracy of detecting results.

The invention discloses a method and a system for checking patrol attendance of security personnel based on mobile phone GPS positioning. The method comprises the following steps: S1) planning a target patrol route of the security personnel; S2) acquiring real-time position information of the security personnel through GPS at the preset patrol time; and S3) comparing consistency of the real-time position information of the security personnel and the target patrol route and alarming if the result is inconsistent. According to the invention, a GPS positioning device is utilized to acquire the practical position of the security personnel at the preset patrol time, and then the alarming is sent when the practical position of the security personnel is inconsistent to a preset position, and a security administrative department is informed of timely knowing the condition of the security personnel deviating from the preset patrol route and position, so that the enhancement for the monitoring degree of the patrol work of the security personnel of the security administrative department is benefited, the effect of enhancing the patrol work of the security personnel is benefited and the security administrative department can timely take measures for settling the abnormal condition.

The invention discloses a network data flow analysis method and a network data flow analysis device based on the cloud security. The method comprises the following steps of capturing a data packet in the network data flow; packaging the captured data packet, and restoring to a message; determining a network protocol corresponding to the restored message; and analyzing the restored message according to the network protocol corresponding to the restored message. According to the technical scheme, the packaging is carried out first, the scattered data packets are regrouped into a significant message, then the protocol analysis is carried out, the analysis can be carried out specifically, and the analysis efficiency and the accuracy can be improved compared with the existing scheme of gradually scanning the data packets one by one.

The invention discloses an effect-directed target / non-target identification method for androgen disruptors in complicated environment samples. The effect-directed target / non-target identification method includes the steps of pretreatment and primary separation of the samples, androgen activity test, high-throughput separation and preparation of toxic substances, active component scanning based on high-performance liquid chromatography-time-of-flight mass spectrometry, target androgen disruptor screening and non-target androgen disruptor identification and toxicant confirmation by combining mass spectrum, chromatography and toxic characteristics. The effect-directed target / non-target identification method has the advantages that an SPE (solid phase extraction) column and a preparation and separation series method are used for separation, a high-throughput separation and concentration method is optimized by DMSO (dimethyl sulfoxide) serving as a protective agent, target identification of key toxicants is achieved through a target database, and mass spectral characteristic identification based on a high-performance liquid chromatography-time-of-flight mass spectrometry library, chromatographic characteristic identification based on retention time and compound physico-chemical property and toxic characteristic identification based on a molecular dynamics simulation technology are conducted to achieve non-target identification / toxicant confirmation of the key androgen disruptors.

The invention discloses a mechanical arm obstacle avoidance method based on multiple sensors and a mechanical arm. The method comprises the steps that the surrounding information of an target object is obtained on the basis of one or more laser sensors and visual sensors; whether an obstacle exists between the tail end of the mechanical arm and the target object is judged; if the obstacle exists,the position of the obstacle is primarily positioned on the basis of image information; the actual distance information of the tail end of the mechanical arm and the obstacle is collected through theone or more laser sensors; a raster map is built according to the primary positioning result and the actual distance information collected by the laser sensors; and on the basis of the raster map, a path is planned through a method for combining manual potential field and a rapid exploring random tree, and the mechanical arm is controlled to get close to the target object. The obstacle in the environment is detected through combination of the visual sensors and the laser sensors, and the obstacle avoidance precision is high.

The invention provides a water-quality monitoring pre-treatment device. The water-quality monitoring pre-treatment device comprises a Y-type filter, a water inlet ball valve, a water outlet ball valve and a filter screen, wherein a water inlet of the Y type filter is connected with a sampling pump by a water inlet tube, so that primary filtration treatment is carried out on an original water sample collected by the sampling pump; a first outlet of the Y type filter is connected with the water inlet of a sampling cup by a water sample conveying pipeline; the filter screen is arranged at the water inlet of the sampling cup; the filter screen is used for carrying out secondary filtration treatment on the primarily filtered water sample conveyed by the water sample conveying pipeline; a second outlet of the Y type filter is used for discharging the primarily-filtered substances through the water outlet tube; the water inlet ball valve is arranged in the water inlet tube; the water outlet ball valve is arranged in the water outlet tube; water sample flow rate in the water sample conveying pipeline is controlled by regulating opening degree of the water inlet ball valve and / or the water outlet ball valve. Besides, the invention further provides a water quality monitoring system which is used for improving representativeness of a water sample and accuracy of the monitored result.

The invention relates to a method for detecting the vanadium content of desulfurized liquid, belonging to the technique of gas desulfurize purification, which can use light split photometric method to measure the vanadium content of desulfurized liquid. It is characterized in that: it uses light split photometric method to measure the vanadium content of desulfurized liquid, while needing the raw materials as: sodium acetate, tannic acid, vitriol, sodium sulfite, and sodium metavanadate standard solution, and the needd devices comprise: spectrophotometer. The inventive method is simple which only adds one spectrophotometer, and adjusts the wavelength to 590nm, and adjust other parameters as pH valve according to the actual condition, then dilutes the desulfurize solution, draws the standard curve to make regress equation, marks the light adsorption degrees of unknown desulfurize solution, and find the vanadium content according to the standard curve. The invention has high accuracy, 100% recycled rate, and 1% relative error.

The embodiment of the invention discloses a video analysis method and device, equipment and a storage medium. The method comprises the steps of obtaining bullet screen information sent when audience users watch videos; Identifying an emotion tag to which each piece of bullet screen information belongs; Determining the proportion of the bullet screen information belonging to each emotion tag; Calculating a quality evaluation value of the audience user to the video according to the proportion; The problem of inaccurate evaluation and one-sided evaluation caused by evaluation of the video qualityonly through bullet screen information is solved, so that the quality of the video content is further analyzed according to the emotion of audience users on the video, the analysis accuracy of the video content is improved, and the video generating emotion resonance with the audience users can be found indirectly.