35results about How to "Improve misjudgment" patented technology

The invention relates to an obstructive sleep apnea-hypopnea syndrome (OSAHS) screening method and a device. Formant characteristics of snore are adopted as technical indexes for screening OSAHS; Snore sections are extracted by a short-time energy method at first, then LPC (linear predictive coding) is used for modeling, after predicting factors of the LPC are obtained, a first formant frequency of each snore section is obtained by a root extraction method, as ranges of formant frequencies of different modes are different, the different snore sections can be distinguished, all first formant frequencies of all snorers are classified, by the aid of K mean values, and thresholds for distinguishing normal snore and abnormal snore of individuals are determined according to classified results; and then AHI (apnea-hypopnea indexes) in a PSG (polysomnogram) are simulated so that average times of apnea and hypopnea of the snorers within an hour are obtained. According to the OSAHS judgment standard, whether the data section has snore of a normal person or snore of a patient with the OSAHS is judged. The obstruction sleep apnea-hypopnea syndrome screening method and the device are low in cost, high in applicability and simple in operation and are non-contacting.

The invention discloses a method for real-time detecting image dead dot pixel, which is applied in a digital camera to real-time detect dead dot pixel existing in images so as to find out pixels which need restoring when the digital camera acquires images. Through real-time detection, the labor power and labor hour which are spent on the dead spot correction program in the production of digital cameras are decreased, and also the storage space used for storing dead dot pixel information for a digital camera is decreased; more importantly, in the process of detection, dead pot pixel can be judged according to different attributes of images thus dynamically adjusting threshold value; therefore, that what happens to the dead pot pixel can be more accurately mastered thus reaching the effect of improving the quality of images.

The invention discloses a warm air blower control method. The method includes a step of setting automatic control by a user and a step of running automatic control, the step of setting automatic control includes comparing environment temperature, namely, when the user sets temperature, comparing a difference value between the environment temperature Tr and user-set temperature Ts with a specified upper-limit temperature value and a specified lower-limit temperature value, the step of running automatic control includes that a main controller controls a heater and an air blower to be turned on and off according to comparison results in the step of comparing environment temperature to realize different running modes. By the warm air blower control method, a warm air blower is enabled to meet real requirements of the user more timely and more accurately to realize good heating effect, and the warm air blower is enabled to automatically correct misjudgment and misoperation.

The invention relates to detection equipment, a detection method and device and a storage medium. The detection equipment comprises a radar module used for emitting radar waves, receiving echoes returned based on the radar waves, and determining motion parameters of a detection object in a radar wave action range according to emission parameters of the radar waves and receiving parameters of the echoes; an infrared detection module used for detecting infrared change parameters in the detection space; and a processing module connected with the radar module and the infrared detection module andused for determining whether a target object exists in the detection space or not according to the motion parameters and the infrared change parameters and determining the state of the target object when the target object exists so as to obtain a determination result. Thus, by integrating the radar module and the infrared detection module in the same equipment and combining the functional advantages of the radar module and the infrared detection module, whether a human body exists or not can be accurately judged, the state of the human body can be determined, and better detection service is provided.

The invention provides a touch panel which comprises a first base plate, a second base plate, a first sensing layer, a second sensing layer and a plurality of clearance objects, wherein the second base plate and the first base plate are arranged in parallel and are opposite to each other up and down. The first sensing layer is configured on the first base plate, and is positioned between the first base plate and the second base plate; and the second sensing layer is configured on the second base plate. The clearance objects are positioned between the first sensing layer and the second sensing layer, and comprise a plurality of removable first clearance objects and a plurality of second clearance objects fixed on the second base plate, wherein each first clearance object is surrounded by part of second clearance objects.

The invention discloses a falsifying-preventing electronic strip seal. The falsifying-preventing electronic strip seal is composed of an insertion bolt and a bolt base. A bolt body of the insertion bolt comprises a rod body, a cap head extending outwards from the top end of the rod body, a vertically-through containing hole and a first clamping structure located at the bottom edge of the rod body. A counterbore with the inner diameter larger than or equal to a bolt hole is formed in the top edge of the containing hole in an outward-extending mode. A circuit board is installed in the containing hole and comprises a circuit wire, a lower contact, an upper contact and an RFID chip, wherein the lower contact and the upper contact are connected with the circuit wire, and the RFID chip is installed on the circuit wire. A third buckling part larger than the containing hole is installed below the containing hole, an upper elastic sheet abutting against the bottom end of the counterbore and an upper elastic sheet abutting against the bottom face of an upper cover are installed at the top end of the containing hole and used for elastically suspending the circuit board in the containing hole, and the upper contact is controlled to be incapable of being electrically connected with the upper cover; however, when the first clamping structure is inserted into the bolt base and can not be separated, the RFID chip can be electrically connected with an antenna inside the bolt base through the lower contact, and the upper contact can be electrically connected with the upper cover to enable the bolt body to form an external antenna of the RFID chip.

The present invention belongs to the technical field of shaft forging flaw detection, and discloses a rapid elimination method for a shaft forging flaw detection phantom wave. The rapid elimination method is based on phantom wave production factor analysis, is obtained through practice summarization, and comprises the following three operation parts: adjusting the repetition frequency of a flaw detection instrument, changing the frequency of a probe, and increasing a detection sound path, wherein ultrasonic wave flaw detection is performed again through one of the three operation parts, it is determined that the wave F is a phantom wave if the wave F disappears, and it is determined that the wave F is a defect wave if the wave F still exists. According to the present invention, the detection can be performed by using any one of the three detection methods if the appearing signal wave F being the phantom wave or the defect wave cannot be determined, such that the problem of the easy misjudgment of the conventional method is improved, the unnecessary loss is reduced, and the flaw detection accuracy of the shaft forging is improved.

The invention discloses a three-dimensional salient object detection technology of a multi-attention guided neural network. A double-flow neural network is constructed in a two-input mode, and an attention mask graph is generated by using two pieces of information, namely RGB and a disparity map; wherein the mask pattern comprises information of two modes, so that the two modes complement each other, and the principle problem of the disparity map can be avoided; the position information of the object can be quickly determined while the program running speed is increased; secondly, decoding isperformed by using down-sampling coding and up-sampling, and pixels of the image at detail positions are recovered step by step, so that the final accuracy is improved, and the result is closer to thecondition of a real human observation object. The negative effect brought by the principle problem that the pixel is still the highest when the object closest to the disparity map is a non-significant object is solved.

The invention provides a method for improving electricity larceny prevention analysis accuracy based on real-time metering. The method comprises the following steps: 1, preliminarily judging whether an electricity larceny behavior exists in a user electricity meter or not through voltage real-time metering; 2, preliminarily judging whether the electricity larceny behavior exists in the user electricity meter or not through current real-time metering; and 3, judging whether the electricity larcenybehavior exists in the user electricity meter according to the change of real-time metering of the user electric quantity. According to the invention, 1) the total metering value and the data acquired by each user for electricity utilization are obtained based on the unified time reference,and are subjected to difference correlation analysis of long-time sequence, so that the situation of misjudgment or incapability of judging electricity larceny caused by a relatively large line loss error in an electric energy difference method is improved; and 2) the electricity larceny behavior can be detected without additionally adding hardware equipment on the basis of a voltage intelligent algorithm, and the cost is low.

The embodiment of the invention relates to a method and a device for carrying out trembling wave identification on electrocardiosignal data. The method comprises the following steps: acquiring first electrocardiosignal data; performing electrocardiogram data fragment division processing on the first electrocardiogram signal data to generate a plurality of pieces of first fragment data; according to all the first fragment data, performing maximum slope statistical processing to generate first electrocardiogram slope data; carrying out sampling point amplitude overrun frequency statistical processing on each piece of first fragment data to generate first fragment overrun frequency data; performing sum calculation processing on all the first segment of overrun frequency data to generate first electrocardio overrun frequency data; according to the first electrocardio slope data and the first electrocardio over-limit frequency data, a first corresponding relation table reflecting the corresponding relation of the electrocardio slope range, the electrocardio over-limit frequency range and the chattering wave probability is inquired, and first chattering wave probability data is generated. According to the embodiment of the invention, the signal identification precision of the electrocardiograph monitoring equipment can be improved, and the conditions of misjudgment and missed judgment are improved.

The invention relates to a cleaning robot and a material identification method thereof, which is suitable for cleaning robots. The cleaning robot is pre-set with a threshold comparison table and comprises a plurality of comparison thresholds, wherein each comparison threshold defines an interval of accumulated energy information accumulated by the moving plane of different materials within a predetermined period of time. The cleaning robot generates accumulated energy information during the movement, so that the cleaning robot can find the comparison threshold corresponding to the accumulated energy information from the threshold comparison table, and then determines the material of the moving plane contacted by the cleaning robot.

The invention discloses an emergency lane illegal occupation detection method and device, equipment and a medium, and relates to the technical field of image recognition, and the method comprises the steps: employing a trained target detection model to detect a to-be-detected monitoring image, so as to obtain a target detection frame corresponding to a vehicle target in the to-be-detected monitoring image; determining a mass point coordinate of the vehicle target based on the slope of a lane in the to-be-detected monitoring picture and the current position information of the target detection frame; and determining whether the vehicle target illegally occupies the emergency lane or not based on the position relationship between the mass point coordinates and the emergency lane area in the to-be-detected monitoring picture. Through the file, the problem of poor usability in the detection process can be avoided, and the problem that the target on the common lane is misjudged to the emergency lane due to the visual angle can be avoided. Therefore, the problems of misjudgment and poor usability during illegal occupation detection of the emergency lane can be improved.