201results about How to "Reduce the probability of misjudgment" patented technology

The invention discloses a method distinguishing identities of access users, and a device. The method comprises acquiring a plurality of behavior characteristic information of a current access user, acquiring a legal characteristic probability value and an illegal characteristic probability value of each behavior characteristic information in a sample database, performing a condition probability calculation according to the legal characteristic probability value and the illegal characteristic probability value of each behavior characteristic information to obtain a legal access probability value and an illegal access probability value of the current access user, and comparing the legal access probability value of the current access user and the illegal access probability value of the current access user to determine the current access user to be a legal access user or to be an illegal access user. Through adoption of the method and the device, the technical problem of poor accuracy of the fact that the user is determined to be the legal user or the illegal user in the prior art is solved, the effect that the user is accurately determined to be the legal user or the illegal user is achieved, and then the missing judgment rate and the misjudgment rate for the illegal user are reduced.

The invention relates to an online safety detection prewarning device of an elevator. The device comprises a car prewarning device and a machine room prewarning device, wherein the car prewarning device consists of a first sensing unit, a first acquiring unit, a first processing unit and a first near field communication unit; the machine room prewarning device consists of a second sensing unit, a second acquiring unit, a second processing unit and a second near field communication unit; the first near field communication unit and the second near field communication unit are in wireless communication; and the second processing unit is remotely communicated with a remote platform through a remote communication gateway. The invention further discloses an online safety detection prewarning method of the elevator. The device adopts a probability neural network algorithm; and the algorithm enables measured characteristic parameter thresholds to continuously self-learn, self-adapt and self-adjust along with such conditions as elevator ageing degree, maintenance condition and load magnitude, so that the wrong judgment or judgment missing probability is largely reduced, and the long-time use reliability and accuracy of a prewarning function of the device are improved.

The method sets up a junk contact mode data base and comprises: A) deciding if the message to be sent carries the contact mode; if yes, then entering the step B; otherwise, entering step D; B) extracting the contact mode carried in the message to be sent; deciding if there is a same contact mode in the junk contact mode database; if yes, entering the step C; otherwise, entering step D; C) rejecting the sending of the message to be sent, and ending the current process flow; D) sending the message to be sent. The invention also reveals an information filtering apparatus.

The invention discloses a voice detection method and device, electronic equipment and a storage medium. Deep semantics of a voice expression can be understood, implicit violation content contained invoice can be detected, the missed judgment probability is reduced, and the recognition precision and recognition efficiency are improved. The method comprises the following steps: carrying out voice recognition on to-be-detected voice to obtain a target text; and based on a trained text classification model, extracting a text feature vector of the target text, and acquiring a prohibited category corresponding to the target text according to the text feature vector, wherein each training sample in a first training sample set for training the text classification model comprises a text sample anda prohibited category corresponding to the text sample.

The present invention discloses an image dead pixel detection and processing method. The method is mainly used for processing dead pixels in an image, can well process dead pixels at the edge of the image, and can well process multiple dead pixels. Firstly, according to a condition of a hardware image sensor, a condition that the method processes most continuous dead pixels is selected; then the dead pixels are processed row by row through iteration, a pixel before the current pixel is a reliable pixel, the current pixel is the center, and a matrix of surrounding 5*5Bayer pixels is established; according to whether the center point is a G pixel, nine different pixels are selected as a reference; according to an image noise threshold, whether the current pixel is a dead pixel is determined; according to difference values between the eight surrounding pixels and the current pixel, whether the pixel is at the edge of the image is determined; and according to different direction edges at which the dead pixels are, different reliable pixels are selected as substituting pixels, and a value is assigned to the current dead pixel so as to correct the dead pixel.

The invention discloses a multi-target life detection method based on a radar signal and a detection radar. Time accumulation is carried out on a pre-processed echo signal in a slow time direction toacquire a first echo signal. The first echo signal is subjected to inflection point envelope extraction to acquire a second echo signal. The average value of all amplitude signals other than M mark amplitude signals in the second echo signal is acquired. If the ratio of the mark amplitude signals to the average value is greater than a threshold value, a life target is at the radar detection distance corresponding to the mark amplitude signals. According to the life detection method provided by the invention, a normalization method is used to normalize the radar signal amplitude in the fast time (distance) dimension; the problem of energy attenuation caused by the increase of the radar signal with distance in the process of transmission is solved; the missed rate of a far-end target is effectively reduced; and the method of two inflection point envelope extraction is used to improve the multi-target axial recognition accuracy.

The invention relates to a vehicle electrical-apparatus sound control method based on a command-word table, which belongs to the technical field of vehicle electronic control. The method comprises the steps of generating a sound word table and a noise word table, making use of a sound identifying engine to record the sound of the driver, combining the sound word table with the noise word table, then loading the combined word table to the sound identifying engine and selecting the word with the pronunciation similar with the sound of the driver from the sound word table, identifying the speech intention of the driver according to the word output by the sound identifying engine, and the identified result is the determined electrical-apparatus control command, converting the electrical-apparatus control command into a CAN bus control signal and outputting the control signal to the CAN bus to control the electrical-apparatus. Due to the use of the control method of the invention, during the vehicle-running process, the controlling performance for the noise is stronger, the error operation is reduced considerably, and the identification precision is improved.

The invention relates to a patrol robot, comprising: a car body, a driving motor, a steering driver, a driving wheel, a camera mechanism and a control module; At the top, the camera is connected to the lifting assembly. The lifting assembly includes a lifting motor and a lifting rod. The lifting motor is set in the storage cavity. There is a lifting hole on the top of the car body. The lifting hole communicates with the storage cavity, and the lifting rod is movable through the lifting hole. Inside, the lifting motor is driven and connected with the lifting rod, and the lifting rod is connected with the camera. The lifting rod moves in the vertical direction under the drive of the lifting motor, so that the camera can shoot the platens of different heights, and the lifting rod is restricted by the lifting hole to avoid the horizontal deviation of the lifting rod, so that the movement of the lifting rod is more accurate. In order to be stable, improve the clarity of the image captured by the camera, and effectively reduce the misjudgment rate and missed judgment rate of false throwing and returning.

The invention relates to a scanning and transceiving integrated remote sensing measuring device and method for motor vehicle exhaust based on TDLAS-based (tunable diode laser absorption spectroscopy-based). Emergence wavelength of lasers covers target absorbing peaks of gas under test through temperature and current control of the lasers; a free light output multi-wavelength coupling unit provideshigh coaxiality or superposition of multiple beams; the coupled beam gains transceiving integration and beam scanning through an off-axis parabolic mirror and a rotating hexahedral reflector; the rotating hexahedral reflector enables the coupled beam to scan exhaust emission plumes and reach strip reflecting stickers on a ground; the transceiving integrated structure focuses light, reflected by the reflecting stickers, to a detector; a detector signal is received and processed by an acquisition processing unit so as to exhaust emission information; the device also includes a wavelength locking unit that locks wavelength of each laser in real time; frequently calibrating a system is not required. The device has high sensitivity and high spectral definition, and is suitable for real-time efficient measurement of exhaust emission.

The invention discloses a signal processing method for a borborygmus signal monitoring system. According to the method, after borborygmus data are quantized into digital signals with sampling rate of Fs by a microphone, the digital signals are treated based on the self-adaptive filtering algorithm and sectioned according to window length, the signal frequency spectrum of each signal section is solved by means of the fast Fourier transform algorithm, the signal frequency spectra are firstly applied to intensity detection, then subjected to correlation-based detection and analysis with a median frequency spectrum model within the frequency band to be tested, and finally sent to a comparator to be compared with a reference value, only the signals with correlation coefficients high enough can be shown in a data window as borborygmus signals and presented by the output result 1, and 0 is output otherwise. On this basis, RIM, IMF and AMT are also provided to be used for evaluating the intestinal tract activity condition of a testee. The method is used for the wearable human body borborygmus signal monitoring system and can be implanted in equipment hardware or a monitoring server in hospital, and the position of a borborygmus of the testee can be marked according to borborygmus signals input into the system, so that technical support is provided for a doctor to offer reasonable postoperative rehabilitation nursing to abdominal surgery patients.

The invention is suitable for crystalline grain sorting and arranging technology, providing a quick sorting and arranging machine of crystalline grain and a quick sorting and arranging method of the crystalline grain. A loading tray to be classified is moved to an image shooting device for shooting images to obtain the current position data of the crystalline grains, wherein plural crystalline grains are arranged on the loading tray, and all crystalline grains are detected by a detection machine so as to ensure the electrical characteristic of each crystalline grain and respective position; the current position data of the crystalline grains is compared with prestored position data to confirm that the crystalline grain of each crystalline grain characteristic is put on the position of the loading tray to be classified. Then, the loading tray to be classified is moved to an adsorption position to ensure that a picking and placing device adsorbs the crystalline grain to put the crystalline grain on the corresponding classified loading tray under the monitoring of a loading monitoring device with high resolution and small-range shooting; meanwhile, one piece of crystalline grain monitoring equipment monitors the crystalline grain placing position in real time; multiple groups of detection equipment receptively divide labor and operate, which not only improves working efficiency, but also reduces the problem of operation overloading caused by single detection equipment.

The present invention is applicable to the field of the smart device, and provides a danger detection and warning method and a danger detection and warning smart device. The danger detection and warning method comprises: detecting whether the motion state of a vulnerable group is changed or not; detecting sound around the vulnerable group when the motion state of the vulnerable group is changed; determining whether the similarity of characteristics of detected sound and characteristics of a preset sound exceeds a preset similarity threshold or not, and/or, determining whether the decibel of the detected sound is larger than a preset decibel; recording detected sound and obtaining position information where the vulnerable group is located when the similarity of characteristics of detected sound and characteristics of the preset sound exceeds the preset similarity threshold, and/or, when the decibel of the detected sound is larger than the preset decibel; and sending recorded sound and the position information where the vulnerable group is located to an assigned smart terminal. The embodiment of the present invention is helpful in doing saving the electric quantity of the danger detection and danger warning smart device.

This invention discloses a method and a system for filtering communication information in the communication system, in which, the filter system includes: a background management module used in providing interfaces of input devices, a database used in storing filter conditions and varied information, a front filter module including a matching module and a filter module. Said method includes: presetting a filter condition and its varied information, matching contents of the communication information transmitted currently and filtering it if the content of the communication information matches to the filter condition or its varied form.

The invention relates to a cloud monitoring based network information security identification method and a cloud device. The method comprises that the cloud device receives security identification requirements carrying network information and client identification, which are uploaded by a client; the cloud device inquires whether the client identification is in a preset identification blacklist; the cloud device feeds back to the client that the identification result is safe false identification information if the client identification is in the preset identification blacklist; the cloud device performs safety identification on the network information and feeds back the identification result to the client if the client identification is not in the preset identification blacklist. According to the method and the cloud device, false identification information is fed back to the client in the preset identification blacklist, so that hackers of the client think that the uploaded malicious network information safety identification result is safe, the problem that identification determination missing of malicious network information of the cloud device caused by the fact that the hackers bypass the cloud device safety identification method in the prior art can be effectively solved, and the determination missing rate of malicious network information is reduced.

The invention relates to the field of automobile safety, and discloses a triggering method of automatic emergency braking system based on prediction, comprising methods for triggering primary alarm, secondary alarm and emergency brake, and meanwhile, the invention also discloses a movement locus predication algorithm of the current automobile and a front automobile. According to the invention, three stages after the system is started is divided according to three stages, namely, primary alarm, secondary alarm and emergency brake, according to the operation mode of the automatic emergency braking system, and meanwhile, a deceleration curve of the automatic emergency braking system is designed according to the difference of the secondary alarm manner of the system and the response characteristic of the vehicle brake system; and in the actual operation process, triggering of various difference stages of the automatic emergency braking system is realized by performing predication computation on the current vehicle and the front vehicle, and therefore, the triggering method has the advantages of being accurate and reliable in triggering, and low in erroneous judgement and missed judgement rate.

The invention relates to a power grid data association method based on an address matching technology. The method achieves the aim that repair and complaint work orders without client numbers are associated to basic file information of clients through an address fuzzy matching technology. The power grid data association method comprises the following steps that pre-processing is conducted; repairaddress information of the clients is received and stored as text information; structured address information, special characters and Arabic numbers are deleted; Chinese address character strings aresubjected to editing distance calculation; a Chinese address character string of a user corresponding to a minimum calculation result is determined as a repair address of the user. Compared with the prior art, the power grid data association method has the advantages that a mode for achieving structured data association by using non-structured data association is provided, and fusion analysis of trans-disciplinary data is achieved; an address matching degree is calculated based on a minimum editing distance algorithm, adoption of Chinese word segmentation is avoided, and the misjudgment probability is lowered.

The invention discloses a power distribution network bad data identification method based on mass tags, and belongs to the field of power system calculation and analysis. Under the condition that multi-source data are acquired, corresponding data mass scores are evaluated through constraint rules of six influence factors such as voltage amount measurement U, current amount measurement I, active amount measurement P, active electricity measurement and reactive electricity measurement, mass tag values are calculated through the mass scores measured by the amounts after evaluation is finished, and bad data are identified according to the calculation result. The defect that in a traditional data identification method, local data errors are shared to all nodes is overcome, redundancy measured data of an upper voltage grade and a lower voltage grade are comprehensively utilized, in combination with the mass tags, the bad data are identified, and the quality of power distribution network virtual measurement and state estimation input data is improved.

The invention relates to a medical data duplicate checking and associating method and system. The method comprises the following steps of: (1) extracting core data items in to-be-processed medical data; (2) classifying the core data items; (3) respectively carrying out preliminary screening on the data items in an exclusion array and a fuzzy array; (4) carrying out deep screening on the data items in the core data items; (5) setting a threshold value M2 of a suspected duplicated data similarity and/or a threshold value M3 of suspected associated data; and (6) after artificially checking and judging the suspected duplicated and/or associated data, inputting the data which is judged as non-duplicated data into a medical database, and endowing the data which is judged to be associated with one or more corresponding association labels. Compared with the prior art, the method and system provided by the invention have the characteristics of being low in missed judging rate, low in wrong judging rate and high in duplicate checking efficiency, and do not have high requirement for the profession degree of artificial checking, so that the operation costs of the duplicate checking and associating are remarkably reduced.