56results about How to "Reduce the number of false alarms" patented technology

This invention discloses a test method for wafers including: utilizing an optical process unit in an AVI tool to get an optical image of a wafer, carrying out grey computation to the image and dividing it to several zones, selecting a subarea standard chip, finally, carrying out contrast of grey values to other chips in different areas with the standard one to test defect chips, in which, the mode of partition is decided according to the state of products, a subarea can be any area in the image and the standard chip may be one existing in the chip or a virtual chip model got by the AVI tool, and a judgment standard should be designed in advance before contrasting grey values.

The invention provides a pipeline leakage feature vector extraction method based on an improved wavelet packet and relates to the technical field of signal feature extraction. The method comprises the first step, the second step and the third step of acquiring a signal and conducting denoising on the signal, the fourth step of adopting an improved wavelet packet algorithm to reconstruct a single-band signal, and the last two steps of extracting a feature vector. In order to overcome the shortcomings of a traditional wavelet packet algorithm, the improved wavelet packet algorithm is disclosed, and a signal prolongation mode and a signal processing mode after convolution are combined to improve the wavelet packet algorithm. A parabola prolongation mode is adopted in the improved algorithm, and FFT and IFFT are adopted in the signal processing mode after convolution. Test results show that the phenomenon of redundant frequency components generated during reconstruction of the single-band signal can be overcome by the improved algorithm, a leakage feature signal is accurately reconstructed, and the feature vector is accurately extracted. The pipeline leakage feature vector extraction method greatly increases detection accuracy, reduces the false alarm rate and also lays a foundation for improving follow-up locating precision.

The invention discloses an automobile transmission rack endurance test operation state monitoring method. The automobile transmission rack endurance test operation state monitoring method includes the following steps that firstly, all sorts of signals are obtained; secondly, a rotation speed impulse signal, a torque voltage signal, a vibration acceleration signal and a gear acceleration signal are processed and then input into a data collector to be converted into digital signals which are input into a computer; thirdly, all monitoring zones are divided; fourthly, the computer performs angular domain recollection by means of the input signals to obtain an order spectrum; fifthly, self-learning reference spectral line generation of all the monitoring zones is performed; sixthly, the real-time order spectrum obtained through angular domain recollection is compared with a reference spectrum, if a preset value is exceeded, an alarming order is started. According to the automobile transmission rack endurance test operation state monitoring method, zone division is performed on gears, rotation speeds, torque and oil temperature parameters, self-learning can be performed on all the monitoring zones to generate a reference order spectral line. The problem that a transmission is complicated in operation conditions and difficult to monitoring is solved. The frequency of faulted alarming of a monitoring system is reduced and monitoring accuracy of the system is improved.

The invention relates to the technical field of electrocatdiogram monitoring. An automatic alarm system for the electrocatdiogram monitoring comprises a service terminal. The service terminal is provided with a data processing and window dividing unit, a heart rate chain alarming unit, a body gesture judging unit, a motion degree judging unit and an electrocatdiogram processing unit, wherein the body gesture judging unit, the motion degree judging unit and the electrocatdiogram processing unit are respectively connected with the data processing and window dividing unit and the heart rate chain alarming unit. The invention also provides an alarm method of the automatic alarm system. Under the condition of referring a motion state and combining the factors of the heart rate with the body gesture, a threshold value of heart rate alarming is adjusted, the accuracy of the electrocatdiogram monitoring alarming is increased, the times of error alarming is reduced, the response efficiency of doctors is increased, and the expense of medical mechanisms and the use cost of a user are saved.

The invention belongs to the field of metallurgy continuous casting production process control, and more specifically relates to a crystallizer bleed-out forecasting system based on logic judgment. According to the system, a hardware component comprises a combined type crystallizer, thermocouples, bleed-out forecasting system computers, a PLC, an interchanger, thermocouple compensation cables, an audible and visual alarm, a relay and a remote station, and further comprises a data acquisition unit, an automatic bleed-out data compression storage unit, a self-learning and off-line analysis unit and an alarm and execution unit, wherein the bleed-out forecast system computers comprise a client-side computer and a server computer, the client-side computer is used for monitoring and displaying bleed-out data, and the design of embedding rows and line spacing of the thermocouples enables the time required to propagate a thermocouple interval in the transverse direction and in the longitudinal direction of bonding tearing openings is the same. The system has the advantages that the timely and accurate forecast for bonding bleed-out can be realized, the smooth continuous casting production can be ensured, and the casting blank quality can be improved.

The invention belongs to the field of control of metallurgical continuous-casting production techniques, and relates to a crystallizer breakout prediction method based on logical judgment. The crystallizer breakout prediction method based on logical judgment comprises the following steps that firstly, logical judgment rule parameters are read; secondly, data are acquired and processed; thirdly, temperature rise of thermocouples is detected; fourthly, the temperature rise rate of the thermocouples which meet the temperature rise judgment condition is detected; fifthly, temperature change delay of the three thermocouples around the thermocouple TC (i, j) with abnormal temperature rise is detected; sixthly, the seven steps is conducted after detection of continuous temperature fall of the thermocouple TC (i, j) or detection of the temperature change delay of the thermocouples around the thermocouple TC (m, n); seventhly, temperature inverted detection of the thermocouples is conducted; and eighthly, whether a sticking breakout alarm needs to be given out or not is judged according to the alarm shielding condition. By adoption of the crystallizer breakout prediction method based on logical judgment, timely and accurate prediction of sticking breakout is achieved, smooth proceeding of continuous-casting production is ensured, and the quality of cast blanks is improved.

The invention discloses an intelligent monitoring alarm system interface of a thermal power plant and an implementation method of intelligent monitoring alarm system interface. Based on a big data technology, data analysis is performed, an algorithm model is utilized to establish a model, original monitoring tens of thousands of equipment characteristic values are fused into one interface to perform monitoring operation, manual data analysis is replaced, early warning and monitoring are performed on operation states of the equipment, the intelligent monitoring system based on DCS data is developed by using technologies such as the Internet and artificial intelligence, centralized monitoring, management, early warning/diagnosis and analysis of the generator set are realized, and the system interface has the beneficial effects that operation data analysis and abnormality judgment of monitoring personnel are reduced, and labor force is released; fault symptoms are warned in advance, and the safety of the unit is improved; early warning reasons are traced, causes are analyzed, alarm reasons are analyzed, and early warning factors are intelligently searched; the false alarm frequency is reduced; and requirements of data acquisition, data processing, model training, model calculation, model management integration and the like are met.

The invention discloses a railway wagon swing bolster spring jump-out fault image recognition method, relates to the technical field of freight train detection, and aims to solve the problem of low detection accuracy of a manual detection mode in the prior art. The method comprises the following steps: 1, obtaining a passing wagon image, and carrying out the coarse positioning to obtain a coarse positioning gray image set; 2, marking the coarse positioning grayscale image set to obtain a marked image set; 3, training a deep learning network model by using the marked image set and the originalimage; and 4, segmenting the swing bolster spring by using the trained deep learning network model, and judging whether a swing bolster spring jump-out fault occurs or not according to a segmentationresult. According to the invention, manual detection is replaced by an image automatic recognition mode, vehicle faults can be automatically identified and alarmed, the operation standard is unified,the influence of different experience, understanding and cognition degrees of detected vehicle personnel is avoided, and the detection efficiency and accuracy are improved.

The invention discloses an optimal target detection scale generation method in people flow rate statistics. The generation method comprises the following steps: obtaining a video source and initializing a statistical target queue people_list to be null; reading current frame image of the video source, converting the image into a grayscale image Gk, and initializing a detection target queue head_list to be null; dividing the grayscale image Gk into n areas; carrying out target detection on the grayscale image Gk, updating the people_list by utilizing detection targets and track information and taking newly-appeared unassociated detection targets as new targets and add the new targets to the people_list; after processing a certain number of video frames, calculating a target detection scale distribution histogram of each area according to the current position information of all statistical targets in the people_list, and calculating the optimal target detection scale of each area by utilizing target detection scale distribution histogram information. The optimal target detection scale generation method promotes high detection rate and low false alarm rate in the people flow rate algorithm based on videos.

The invention relates to a traffic big data suspect vehicle inspection deployment and control method and device. The method includes the following steps that: a suspect vehicle blacklist and suspect vehicle characteristic information are determined by means of a suspect vehicle discovery method; when a vehicle of which the license plate number is same as the license plate number of a suspect vehicle appears, and whether the vehicle is a suspect vehicle or normal vehicle is judged according to the basic information and characteristic information of the suspect vehicle, if the vehicle is the suspect vehicle, an alarm is issued, and the suspect vehicle will be inspected and intercepted; an intelligent traffic signal machine near the suspect vehicle can perform coordinated control, so that time for the suspect vehicle to pass through an intersection can be prolonged; and if the vehicle is a normal vehicle, no alarms are issued. With the method and device of the invention adopted, the discovery rate of vehicles with fake license plates and the success rate of the inspection of the vehicles with fake license plates can be improved, and the waste of police resources can be avoided to a greatest extent.

The invention discloses a method for using a scanning machine program to detect wafers according to floating threshold values. The method includes the first step of collecting the relations between front-layer processing measurement data of each wafer and the type of noise defects at a current site and the relations between the front-layer processing measurement data of each wafer and the number of the noise defects at the current site and fitting the function relationship between the threshold values and parameters, the second step of allowing a scanning machine to call front-layer processing measurement parameters of each wafer when a product reaches a scanning site, and the third step of substituting the parameters into a fitting function in the first step to obtain one or more threshold values of each wafer, leading the threshold values into the scanning program and scanning each wafer according to the threshold values of each wafer by using the scanning program sequentially. According to the method, the related front-layer processing data of the wafers in the manufacturing process are collected through the scanning machine, then one or more threshold values are obtained through the fitting function calculation of the scanning machine, defect scanning is carried out according to the floating threshold values so as to eliminate influences of noise defects caused by tiny processing difference between the wafers, and consequently the number of the wafers with specifications exceeding false-alarm can be reduced.

The embodiment of the invention provides a service alarm method and device and a storage medium. The service alarm method comprises the steps obtaining values of multiple service indicators within thecurrent cycle according to a preset alarm strategy, wherein the multiple service indicators respectively correspond to different attributes of a service; and outputting alarm information in the casethat the values of the multiple service indicators respectively exceed alarm thresholds corresponding to the multiple service indicators. According to the method, the multiple indicators of the service are monitored according to the alarm strategy, and the alarm is only given when all of the multiple service indicators meet the alarm thresholds, thereby better conforming to alarm requirements of the operation and maintenance personnel, and reducing the number of false alarms.

The invention discloses a non-coherent radar image background modeling method based on a normal distribution function. The method is used for radar image processing and target detecting. According to the method, a time domain background pixel sample set is built by utilizing the grey levels of pixels in a non-coherent radar image background image sequence, the probability distribution features of the grey levels of the background pixels are described with the normal distribution function, and the background pixel area is distinguished to be an airspace area, a fixing target interior area and a fixing target edge area according to the mean value and the variance. In a background model, the pixels of the airspace area are calibrated to be zero, the pixels of the fixing target interior area are one, and the pixels of the fixing target edge area are the decimals between zero and one. The background model is built based on the time domain probability distribution features of the grey levels of the pixels in non-coherent radar images and is the important foundation of low-altitude airspace radar target detection, the detection capacity for targets in the airspace area in the non-coherent radar images can be improved effectively, and meanwhile the frequency of false alarms in the fixing target edge area is reduced.

The invention provides a rescue device alarm method and system and electronic equipment, and relates to the technical field of rescue devices. The rescue device alarm method comprises the steps of judging whether or not a first motion signal transmitted by a motion detection device is received within a first preset period of time; if yes, determining the state as the normal state; if not, determining the state as an early-warning state, and controlling the motion detection device to perform pre-alarming by transmitting early-warning signals to the motion detection device; judging whether or not a second motion signal transmitted by the motion detection device within a second preset period of time; if yes, determining the state as the normal state; if not, determining the state as the alarmstate, and sending the alarm signal to a terminal. The technical problem that in the prior art, an existing rescue device gives a wrong alarm, and therefore the wrong alarm frequency of the rescue device is excessively large is solved.

The invention provides a snow depth measurement method based on image contrast. In a condition in which an abrupt change happens when the snow depth value is measured by laser, a camera is started for capturing, whether the abrupt change is caused when a covering intrudes the measurement area is judged through an image contrast method, and if yes, a foreign object is eliminated. Through the detection method provided by the invention, interference to measurement of the snow depth value by the covering can be effectively eliminated, the calculation real-time performance is good, combination with the existing video monitoring system is facilitated, and the practical value is strong.

The invention discloses a fault alarm method and device. According to the method, after an indicator diagram of an oil pumping well is obtained, if the situation that a fault exists in the oil pumpingwell is analyzed based on the indicator diagram of the oil pumping well, information of the fault is recorded; and when the fault is analyzed from a continuous preset number of pieces of the indicator diagrams of the oil pumping well, alarm information that the fault exists in the oil pumping well is output. The scheme of the method and device can reduce the false alarm rate of the fault and improve the accuracy of fault alarm.

The invention discloses a secondary target screening method based on multi-beam forming, thereby solving problems of large computational load and low detection efficiency in the prior art. The methodcomprises steps: step one, determining a target unit of a clutter sidelobe region; step two, calculating a ground clutter azimuth with the same Doppler frequency as the target unit; step three, determining an azimuth range of the multiple beams and calculating weight vectors of the multiple beams; step four, comparing a difference between an average output power of the multiple beams and an outputpower of the main beam with a secondary detection threshold and determining whether alarming is false alarming; and step five, traversing all target units and storing real target units to complete secondary detection. According to the invention, a plurality of beams are formed in a specific angle range; and whether the alarming is false alarming is determined based on the difference between the output powers of the main beam and the multiple beams, so that the false alarming number is reduced effectively, the detection performance of the moving target is improved, and the computing load is reduced. Therefore, the method can be applied to detection of ground moving targets by the airborne early warning radar in a non-uniform environment.

The embodiment of the invention discloses a method, device and system for constructing a gas concentration prediction model and a storage medium. The method comprises the steps of acquiring the gas concentration of a preset position of a coal mining working face and factors related to the gas concentration in the coal mining working face; and training the pre-constructed neural network model by using the gas concentration at the preset position and factors related to the gas concentration to obtain an optimal neural network model as a gas concentration prediction model. The optimal neural network model obtained by the method can be used for more accurately measuring the gas concentration. More importantly, accurate data of gas can be provided by a supervisory unit, the safety production ofa coal mine is ensured, and the occurrence frequency of gas explosion accidents is reduced.

The invention discloses a target secondary discrimination method after constant false alarm detection, and the method comprises the steps: obtaining a distance Doppler unit passing through a primary CFAR detection threshold, and obtaining a secondary detection unit; calculating clutter steering vectors of clutter points needing to be suppressed in the secondary detection unit, and estimating the amplitude of each clutter point; according to the amplitude of each clutter point, suppressing clutters in the echo data of each distance Doppler unit in a targeted mode, and obtaining suppressed unit data; and performing conventional clutter suppression on the suppressed unit data again, and performing secondary CFAR detection to obtain a target detection result. According to the target secondary discrimination method after constant false alarm detection provided by the invention, the number of false alarms is reduced, and the burden of a subsequent radar data processor is reduced; and the method is small in calculation amount and easy to implement in engineering.

Provided is a method for monitoring a machine tool (10), in particular a material-removing machine tool (10), having the steps of: (a) determining process variable measured values (B(i)) of a processvariable (B) on the basis of a parameter, (b) determining whether the process variable measured values (B(i)) are in a predefined tolerance range (T(i)) which depends on the parameter, (c) if not, outputting a warning signal, and (d) constantly repeating steps (a) to (c), wherein the parameter is a control variable (i), in particular a scalar control variable, which always characterizes progress of the machining process. The invention provides for the control variable (i) to be calculated from the real time (t) and at least one process parameter (O, Atstiii) which characterizes the processingspeed of the machining process.