68results about How to "Troubleshoot technical issues with poor accuracy" 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 discloses an object tracking method and device, a storage medium and an electronic device. The method comprises the following steps: acquiring at least one image acquired by at least oneimage acquisition device; acquiring a first appearance feature of the target object and a first spatial-temporal feature of the target object according to the at least one image; obtaining appearancesimilarity and space-time similarity between the target object and each global tracking object in a currently recorded global tracking object queue; under the condition that it is determined that thetarget object is matched with the target global tracking object according to the appearance similarity and the space-time similarity, distributing a target global identifier corresponding to the target global tracking object to the target object; determining a plurality of associated images acquired by a plurality of image acquisition devices associated with the target object by using the targetglobal identifier; and generating a tracking track matched with the target object according to the plurality of associated images. According to the invention, the technical problem of poor object tracking accuracy in the prior art is solved.

The invention provides a method, device and system for face characteristic information, and relates to the technical field of image identification. The method comprises the following steps: acquiringa to-be-processed first image frame and a to-be-processed second image frame, wherein the first image frame and the second image frame are at least one pair of continuous image frames in a video stream, the first image frame comprises an estimated face frame and the second image frame comprises a detected face frame; and processing an area identified by the estimated face frame of the first imageframe and an area identified by the detected face frame of the second image frame, and tracking the face characteristic information of the first image frame based on a processing result, wherein the face characteristic information comprises a face frame and/or face characteristic points, so that the technical problem that the accuracy of face identification is more difficult when a complicated face is moved by adopting the face positioning tracking method in the prior art is solved.

The invention discloses an emotional information detection method and apparatus for a web text. The emotional information detection method comprises the steps of performing word segmentation processing on the obtained web text to obtain multiple words; acquiring the feature vectors of the multiple words to obtain the feature vectors of the web text; and performing emotional classification processing on the feature vectors of the web text through tuning parameters to obtain the emotional information of the web text. The technical problem of relatively low accuracy of emotional analysis for the web text information in the prior art is solved.

The invention discloses a merchant recommendation method and device, electronic equipment and a readable storage medium. The method comprises the following steps: acquiring click sequence informationof a target user; according to the click sequence information, obtaining a recommendation score of each alternative merchant through a preset recommendation model; based on the recommendation scores of the alternative merchants, obtaining a target recommendation merchant of the target user and pushing the target recommendation merchant to the target user, wherein the recommendation model is a combined model obtained by carrying out combined training on a double-layer recurrent neural network model and a deep neural network model. Therefore, the technical problem that an existing recommendationmethod is poor in accuracy is solved. The beneficial effect of improving the accuracy of recommending merchants is achieved.

The embodiment of the application provides a method and a device for determining the corrosion rate of the bottom of a storage tank. The method comprises: acquiring the working condition parameters ofthe target storage tank; establishing a storage tank model of the target storage tank according to the working condition parameters of the target storage tank; by utilizing the storage tank model, determining a fitting relationship between an acoustic emission signal and the corrosion rate through multiple corrosion experiments; acquiring acoustic emission signal data of the target storage tank,and determining the corrosion rate of the bottom of the storage tank according to the acoustic emission signal data of the target storage tank and the fitting relationship between the acoustic emission signal and the corrosion rate. According to the solution, since the storage tank model approximating the target storage tank is first established, then the accurate fitting relationship between theacoustic emission signal and the corrosion rate is determined through the storage tank model and the acoustic emission signal data of the target storage tank is utilized to determine the corrosion rate of the bottom of the target storage tank according to the above fitting relationship, the technical problem existing in conventional methods that the accuracy of determining the corrosion rate of the bottom of a storage tank is poor is solved.

The invention discloses a wall-hanging stove and a control method and device thereof. The method comprises the steps of detecting whether pairing connection between the wall-hanging stove and first detection equipment is established or not, wherein the first detection equipment is used for detecting the indoor temperature of an indoor space where the wall-hanging stove is located and is in communication connection with the wall-hanging stove; when it is detected that pairing connection is established, receiving the indoor temperature of the indoor space where the wall-hanging stove is located detected by the first detection equipment; and adjusting the output temperature of the wall-hanging stove according to the indoor temperature. By means of the wall-hanging stove and the control method and device thereof, the technical problem that traditional wall-hanging stoves are poor in accuracy of room temperature adjustment is solved.

The invention discloses a wireless broadcast time service calibration method and device, and a computer readable storage medium, and the method comprises the steps: carrying out the initial error estimation compensation of a local crystal oscillator when a signal frame of a starting end is analyzed for the first time, i.e., the signal frame is synchronous with a signal of the starting end, so as to enable the crystal oscillator difference to be preliminarily evaluated and corrected; performing time delay compensation on a synchronous clock at each synchronous moment so as to reduce the influence of time delay on the synchronous clock, determining an accurate timing starting point of a local clock through a signal frame sent by a sending end so as to complete synchronous position locking, wherein the synchronous clock is used for calibrating the local clock, accumulated errors generated by time lapse are eliminated, and the accuracy of long-term timing of the local clock is ensured; andby evaluating the crystal oscillator error in real time and correcting the crystal oscillator error in real time after the synchronous position is locked, the accuracy of short-term timing of the local clock is ensured, and meanwhile, the influence of time service precision caused by environmental influence including multipath in the long-wave transmission process can be reduced.

The invention belongs to the technical field of signal processing and relates to a signal processing method and device for the fault diagnosis of gear box equipment. The method comprises a step of carrying out redundant lifting wavelet decomposition on an obtained discrete signal and obtaining a detail signal and an approximation signal of a current decomposition layer, a step of carrying out singular spectral analysis on the detail signal and obtaining a singular value of the detail signal of the current decomposition layer, a step of calculating a singular spectrum slope of the current decomposition layer according to the singular value of the detail signal of the current decomposition layer, a step of responding to the approximation signal if the singular spectrum slope of the current decomposition layer is smaller than a preset threshold, updating the discrete signal, repeating the above steps, otherwise obtaining a minimum value between the number of layers of the current decomposition layer and 5 to be an optimal decomposition layer number. A collected acoustic emission signal is analyzed to judge an equipment fault condition. A technical problem of poor accuracy of the fault diagnosis of the gear box equipment in the prior art is solved, and the accuracy of the fault diagnosis is effectively improved.

The invention provides a multi-depth camera external parameter calibration method. The method comprises the following steps: shooting the same calibration plate by using a plurality of depth cameras;performing image feature point extraction on the calibration plate image; acquiring image feature points of the calibrated depth camera, and calculating three-dimensional coordinates of the image feature points of the calibrated depth camera; determining an uncalibrated depth camera with the same image feature points; acquiring three-dimensional coordinates of the image feature points of the uncalibrated depth camera with the same image feature points; calculating an image feature point error value between any two depth cameras in all the depth cameras in the adjacent area of each depth camera; and correcting the external parameters of the depth camera, and setting the corrected external parameters as calibration external parameters of the depth camera. The invention further provides a multi-depth camera external parameter calibration device. The accuracy of external parameter calibration of the multi-depth camera is improved.

The invention discloses a test case generation method and device, computer equipment and a storage medium. The method comprises the following steps: receiving test case information sent by a client; if a preset test demand corresponding to the test case information is not obtained from a preset demand database, obtaining a test case, if so, identifying the use case keyword in the test case information, and analyzing and converting the test case information according to the identified use case keyword to obtain a target test case corresponding to the test case information; performing word segmentation processing on the target test case by adopting a preset word segmentation mode to obtain a to-be-analyzed character string, and performing legality analysis on the to-be-analyzed character string to obtain an analysis result; and if the analysis result is that the to-be-analyzed character string is legal, storing the target test case as an executable test case in a preset database. By adopting the method, the technical problem that the test case cannot be generated based on the latest software requirement is solved.

The embodiment of the invention provides an image description generation method and device, and the method comprises the steps: obtaining object data and relation data in a target image; determining theme data of the target image according to the object data and the relation data; calling a pre-trained image processing model to obtain and implicitly deduce high-level semantic information in the target image through bottom-level representation according to the data so as to generate image description corresponding to the target image, wherein the pre-trained image processing model has the capability of abstracting the underlying representation. The method comprises the following steps: abstracting theme data from object data and relational data, and determining image description of a targetimage according to the theme data, the object data and the relational data by utilizing a pre-trained image processing model containing theme nodes; therefore, the image description capable of accurately representing the high-level semantic information in the target image can be efficiently obtained.

The invention discloses a data storage method for a non-volatile storage space in a chip and a trusted chip. The method comprises the step that the non-volatile storage space is created in the chip, wherein properties of the non-volatile storage space at least include parameters used for representing owner data stored in the non-volatile storage space and a storage address range of the owner data,and the storage address range of the owner data is used for representing maximum data length allowing requesting of data from the chip. Through the method, the technical problem that in the prior art, when original data is read from the non-volatile space of the trusted secure chip, the read original data is poor in accuracy is solved.

The invention provides an eyeball identification apparatus. The eyeball identification apparatus is used for performing eyeball identification on a corresponding intelligent mobile terminal. The eyeball identification apparatus comprises an eyeball camera arranged on a body of the intelligent mobile terminal, an infrared fill lamp arranged on the body of the intelligent mobile terminal and on one side of the eyeball camera, and a control chip for performing an eyeball identification operation according to an eyeball image acquired by the eyeball camera. When the eyeball identification apparatus performs the eyeball identification operation, a center line of an image pickup region of the eyeball camera is substantially perpendicular to a preset plane, and an included angle between the center line of a light filling region of the infrared fill lamp and the preset plane where one side of the eyeball camera is arranged is a first acute angle. The invention further provides the intelligent mobile terminal. According to the eyeball identification apparatus and the intelligent mobile terminal which are provided by the invention, by adjusting the image pick-up region of the eyeball camera and the light filling region of the infrared fill lamp, an identification region of the eyeball identification apparatus is larger and closer to the intelligent mobile terminal, so that the accurate eyeball identification operation can be performed closely.

The invention discloses an image fine-grained classification method, system and equipment, and the method comprises the steps: carrying out the preprocessing and normalization processing of an image, inputting the image into a convolutional neural network, and carrying out the feature extraction, thus the conventional image spatial features are considered and the features of an image channel are considered, the fine-grained image classification effect can be improved; then obtaining a proportional relation between different channels through an SENet module, and identifying more fine discriminative regions of the image; adopting a clustering method of metric learning to find a plurality of attention areas in one image, so that the distance of the local features of the same category in the image is further shortened, and the distance of the local features of different categories in the image is further increased; and finally, training the image classification model continuously until the loss function converges to obtain an optimal image classification model, so that the images are classified through the model, and the technical problem of poor accuracy of fine-grained classification of the images in the prior art is solved.

The invention relates to a classification model, and provides a quota prediction method and device and a computer readable storage medium. According to the invention, subsequent insurance amount estimation is carried out only when the risk level is detected to be within a certain range, so that possible negative effects caused by continuous insurance buying of customers with relatively high risk levels are avoided; risk assessment is carried out by selecting the risk assessment model corresponding to the insurance type of this time, so that assessment errors caused by different insurance typesare avoided; a risk assessment model is obtained by using feature engineering and machine learning training, and target features are obtained by using the risk assessment model, so that a risk assessment result finally obtained by the risk assessment model is more accurate; the insurance amount range that the client can understand at present is predicted by synthesizing the previous insurance information and the amount adjustment factor, so that the insurance amount information that can be insured and conforms to the actual situation can be obtained during the current insurance. In addition,the invention also relates to a blockchain technology, and the previous insurance information can be stored in the blockchain.

The invention discloses a candidate character string processing method and device, electronic equipment and a storage medium, and relates to the field of information recommendation. The specific implementation scheme is as follows: acquiring an input original character string based on coordinates of a plurality of input points input by a user; and performing character error correction according toa preset character error correction strategy based on the original character string to obtain a plurality of candidate character strings. According to the technology provided by the invention, character granularity error correction can be carried out on the character string, so that the candidate character string after error correction better conforms to the expectation of a user, the accuracy ofthe obtained candidate character string can be effectively improved, and the input accuracy and the input efficiency of the input method are further improved.

The invention discloses a decontamination verification method and device, a storage medium and an electronic device. The method comprises the steps: obtaining an image set obtained by conducting imagecollection on a decontaminated target vehicle, wherein the target vehicle is about to enter a breeding area to convey goods and materials; obtaining an identification result obtained by performing image identification on an image area corresponding to a target decontamination position of a target vehicle in the image set; and generating a decontamination verification record matched with the target vehicle according to the identification result, wherein the decontamination verification record carries the verification watermark. According to the invention, the technical problem of poor decontamination verification accuracy caused by a manual verification mode provided by related technologies is solved.

The invention belongs to the technical field of gun testing, and in particular, relates to a machine gun recoil energy testing device and method. The machine gun recoil energy testing device comprisesa machine gun cart, a frame, a buffer and travel switches; the machine gun cart is arranged in the frame, and can slide along a rail arranged on the inner side of the frame; the buffer is arranged atthe end of one side of a buffer section of the frame; and three travel switches are arranged at a speed measuring section on the inner side of the frame at the inner side end of the buffer. Through the machine gun recoil energy testing device and method provided by the invention, the technical problem of low machine gun recoil energy testing accuracy in the prior art is effectively solved, the influence on kinetic energy by friction is effectively eliminated, the testing result is more accurate, and the accuracy of machine gun recoil energy testing results is effectively guaranteed.

The invention discloses a dismantling operation control method and device, a storage medium and an electronic device. The dismantling operation control method comprises the steps: in the process thata game application client runs a game task, obtaining a removal notification message, wherein the removal notification message is used for notifying removal of a target object already arranged in thegame task; controlling the virtual character controlled by the game application client to move to the target position of the target object; and under the condition that it is detected that the virtualcharacter enters a demolition area matched with the target position and an observation obstacle is not detected between the virtual character and the target object, controlling the virtual characterto execute demolition operation on the target object, wherein the demolition area takes the target position as a center point. According to the dismantling operation control method, the technical problem of poor accuracy of a demolition control mode provided by related technologies is solved.

The invention discloses a method and a device for testing a synchronous dynamic random access memory. The method comprises the steps that pre-configured first test data are acquired, the first test data are used for testing whether a connection fault exists in a synchronous dynamic random access memory SDRAM or not, and the first test data comprise multiple pieces of test data; writing the first test data into the SDRAM to obtain a first writing result; after a first write-in result is read and a first read result is obtained, comparing the first read result with a first predetermined output result to obtain first statistical data; and under the condition that the first statistical data is greater than a first predetermined threshold, determining that the SDRAM has a connection fault. According to the invention, the technical problem of poor SDRAM testing accuracy is solved.

The invention discloses an air conditioner control method and system, an air conditioner and a storage medium. The method comprises the following steps: obtaining environment parameters of all air supply areas of the air conditioner, and determining the comfort degrees of the air supply areas is determined according to the environment parameters of all the air supply areas; comparing the comfort degrees of all the air supply areas to determine the target comfort degree; and adjusting the angle of wind guide strips to supply air to the air supply area corresponding to the target comfort degree, thus educing the difference value of the comfort degrees of all the air supply areas. According to the air conditioner control method and system, air conditioner and storage medium, the uniformity adjustment of the indoor comfort level is realized.