38 results about "Information distance" patented technology

A light source projection unit projects intensity-modulated spatial pattern light. A time-of-flight distance measurement camera measures a distance to an object on the basis of a time of flight of a modulated component included in reflected light of the spatial pattern light from the object. A spatial information distance measurement camera measures a distance to the object on the basis of spatial information included in the reflected light. A depth synthesizing unit synthesizes measurement results of the distances in the time-of-flight distance measurement camera and the spatial information distance measurement camera to determine a depth value of each pixel position of an image imaged by the time-of-flight distance measurement camera or the spatial information distance measurement camera.

A localization apparatus and method for localizing by use of image information and range data are provided. The localization apparatus includes an image sensor which captures an image and outputs image data, a range sensor which senses a distance and outputs a range sensing signal, an image feature point information processing unit which extracts image feature points from the sensed image and generates image feature point information about the extracted image feature points, a range feature point information processing unit which generates range feature point information about range feature points from the range sensing signal, and a state estimation unit which estimates a state variable including a position of the localization apparatus by using the image feature point information and the range feature point information as an observation value.

The invention discloses a monocular vision road target detection and distance estimation method based on improved YOLOv3, belongs to the field of image processing and computer vision, and is suitablefor intelligent systems of aided driving, road environment perception and the like. The implementation process of the method comprises the following steps: acquiring a road target image marked with road target bounding box information, distance information and category information; constructing an improved YOLOv3 network model based on a Global-Context structure and a cavity convolution pooling pyramid structure; redesigning a loss function of the network in combination with a perspective projection relationship of a camera system and a frame prediction mechanism of YOLOv3; and training and obtaining a road target detection and distance estimation model, and realizing detection and distance estimation of the road target by utilizing the model. Through the monocular vision road target detection and distance estimation method, the distance of the road target can be estimated while the road target is quickly and accurately recognized and positioned in a natural scene, and the monocular vision road target detection and distance estimation method has important significance for practical application.

The present invention provides a discriminant local information distance preserving projection-based speaker confirmation method and belongs to the voiceprint recognition, pattern recognition and machine learning field. According to the method, speech data are obtained in a training phase; the i-vector of each training speech datum is extracted; the i-vector of each speaker is extracted accordingto training speech data corresponding to each speaker; training is performed, so that a discriminant local preserving projection matrix is obtained; during a speaker confirming phase, speech data to be tested are acquired, one speaker of the training speech data is selected, a distance between the speech data to be tested and the i-vector of the speaker is calculated; and if the distance is smaller than a preset distance threshold value, it is determined that the speech data to be tested belong to the speaker; and the confirmation of the speaker is completed. The method of the invention has high applicability, focuses on heterogeneous neighbor points, enhances the discrimination of the easily-confounded speech of speakers, has better distinguishing capability and improves the accuracy of speaker confirmation.

Disclosed is a method and device for aligning at least two digital images. An embodiment may use frequency-domain transforms of small tiles created from each image to identify substantially similar, “distinguishing” features within each of the images, and then align the images together based on the location of the distinguishing features. To accomplish this, an embodiment may create equal sized tile sub-images for each image. A “key” for each tile may be created by performing a frequency-domain transform calculation on each tile. A information-distance difference between each possible pair of tiles on each image may be calculated to identify distinguishing features. From analysis of the information-distance differences of the pairs of tiles, a subset of tiles with high discrimination metrics in relation to other tiles may be located for each image. The subset of distinguishing tiles for each image may then be compared to locate tiles with substantially similar keys and / or information-distance metrics to other tiles of other images. Once similar tiles are located for each image, the images may be aligned in relation to the identified similar tiles.

Provided are methods, systems, and computer program products that generate a plurality of query queues that correspond to a plurality of row-based identifiers corresponding to a plurality of physical memory locations in at least one of a plurality of associative memory networks in response to a semantic-space based query that includes at least one query search element, determine count information responsive to the plurality of query queues and compute information distances between the at least one query search element and ones of associative memory query result elements to determine closeness values corresponding to the information distances.

The present invention relates to a DDoS attack and shanyong event detection method based on stream. The applied stream detection method combines [phi]-entropy improved based on shannon entropy and general entropy and the stream multi-dimensional features. The method comprises the steps of: analyzing multiple types of DDoS attack and shanyong event features, and making multiple types of DDoS attackand shanyong event flow, wherein the made flow can generate particular flow table data of an SDN in the SDN; introducing the [phi]-entropy improved based on shannon entropy and general entropy to increase the information distance between different data to facilitate discovering of attack behaviors as soon as possible; obtaining the multi-dimensional data of the flow table in a switch, such as a protocol type, the flow survival time, the shannon entropy and the general entropy of source / target IPs and the [phi]-entropy, performing feature extraction; classifying different types of DDoS attackflows, the shanyong event flow and the normal flow, namely multi-classification, comparing the detection accuracy of the classification methods such as the SVM and the KNN; and regulating the value ofan adjustable parameter [Alpha] of the [phi]-entropy, and combining an optimum classifier to obtain an optimal multi-classification accuracy. The DDoS attack and shanyong event detection method basedon stream employs the particular flow table function of the SDN and combines the [phi]-entropy to timely detect attacks when the attacks generate and reduce the false alarm rate of the shanyong event.

A system for controlling a station operation mode is provided. The system for controlling a station operation mode, including: a distance information generation unit measuring a distance between a first communication station and a second communication station and generating distance information; a distance information comparison unit comparing the distance information and a predetermined threshold distance; and an operation mode determination unit determining any one of an active mode and an idle mode as an operation mode of the second communication station, based on a comparison result of the distance information comparison unit.

The invention discloses a vehicle-searching system and belongs to the field of wireless communication. The system comprises a vehicle-mounted positioning device, a moveable receiving device and a terminal device, wherein the vehicle-mounted positioning device comprises a radio frequency communication unit I, a satellite positioning unit, a control unit and a power supply unit; the moveable receiving device comprises a radio frequency communication unit II and a transmission unit I; the terminal device comprises a display unit and a transmission unit II; the satellite positioning unit is used for acquiring position information of the vehicle; the radio frequency communication unit I is used for sending the position information of the vehicle to a space; the radio frequency communication unit II is used for receiving the position information of the vehicle; the transmission unit I is connected with the transmission unit II so as to transmit the position information of the vehicle to the terminal device; the position information of the vehicle is displayed by the display unit. The vehicle-searching system can solve the problem of poor vehicle-searching user experience caused by ultra-short Bluetooth transmission information distance and serious signal attenuation of the present vehicle-searching device.

A localization apparatus and method for localizing by use of image information and range data are provided. The localization apparatus includes an image sensor which captures an image and outputs image data, a range sensor which senses a distance and outputs a range sensing signal, an image feature point information processing unit which extracts image feature points from the sensed image and generates image feature point information about the extracted image feature points, a range feature point information processing unit which generates range feature point information about range feature points from the range sensing signal, and a state estimation unit which estimates a state variable including a position of the localization apparatus by using the image feature point information and the range feature point information as an observation value.

The invention discloses an identification method and system for screen operation of an intelligent terminal; a plurality of distance detection modules are arranged below a screen of the intelligent terminal, and the identification method comprises the following steps: obtaining a corresponding relationship between each distance detection module and each piece of coordinate information on the screen; when at least one target object is subjected to air separation operation, detecting different target objects through different distance detection modules, and obtaining distance information betweeneach target object and a screen, wherein the target coordinate information, detected by the distance detection module, of the target object on the screen is in one-to-one correspondence with the coordinate information corresponding to the distance detection module; and according to the target coordinate information and the distance information, identifying the operation behavior of the air operation of each target object on the screen. According to the method and the device, the screen is directly operated in the vertical direction and the horizontal direction in the air operation scene, gesture actions do not need to be preset, the operation is simple, the practicability is high, and the user experience is improved.

A computer implemented machine translation system and method is provided that improves the accuracy of output from one or more machine translation systems by applying one or more data correction routines. A data correction routine is provided that includes information distance analysis of one or more sets of machine translation information to a set of text elements related to the domain and stored to a database. The system and method generate as output corrected text elements related to a meaning intended by a user from whom the machine translation information was captured.

The invention discloses a Web anti-injection method. The web anti-injection method comprises the step A of binding Web application pages and variables, and forming incidence relations among the pages, the variables and operation; a step B of receiving legal data used for training the variables and input by users, and performing calculation and statistic treatment on an information distance of the legal data; a step C of building a recognition model according to statistical information of the information distance of the legal data so as to judge whether input data are legal. The invention further provides a Web anti-injection device corresponding to the method and Web anti-injection equipment with the Web anti-injection device. According to the Web anti-injection method, device and equipment, unknown attack can be defended dynamically, zero-day attack can be resisted, and attackers cannot perform spoofing attack.

The invention provides a character correction method and device, equipment and a storage medium, and the method comprises the steps: obtaining the position information of a click operation of a user on a terminal screen, and obtaining the feature information of the click operation; determining distance information between the position information and each key on a terminal screen; determining transition probability information from a character indicated by the historical click operation to a character corresponding to each key; determining a character prediction result according to the featureinformation, the distance information and the transition probability information, with the character prediction result comprising a possibility score that a character indicated by the click operationbelongs to a character corresponding to each key; and taking the character with the highest possibility score as a correction character corresponding to the click operation. According to the technical scheme, characters can be corrected on the basis of the keyboard layout mode of the soft keyboard on terminals, the situation that correction errors are generated for different terminals is avoided,and the characters clicked by a user can be corrected correctly.

The invention discloses an object volume visual measurement method combining target detection and depth calculation, and the method comprises the following steps: data collection: carrying out the photographing of an image and the collection of distance information through a depth camera; a target detection step: training the measured object by using the YOLO model, and realizing the detection ofthe measured object; an image processing step: processing the detected image to obtain pixel coordinate information required by calculation; and a volume calculation step of obtaining a correspondingdepth value according to the pixel coordinate information obtained in the image processing step so as to measure the size of the object. The depth information (distance information) of a specified pixel point is acquired through image shooting of the depth camera and recognition and sorting of the AI model, and measurement of volume information of an object is realized through image processing andcalculation.

A positioning system, wherein the communication base station has: signal transmitting means for transmitting a base station timing signal and a communication signal; the terminal has: standard timing signal correcting means; total delay information generating means for generating total delay information; and basic information transmitting means for transmitting basic information including the current position information and the total delay information to the information supply device and the like; and the information supply device has: base station position information storage means; basic information receiving means for receiving the basic; distance information generating means for generating distance information indicating a distance between a current position of the terminal and a position of the communication base station; propagation delay information generating means for generating propagation delay; device peculiar delay information generating means for generating device peculiar delay information indicating a delay other than the propagation delay time; and the like.

Embodiments of the invention provide a method and a device for querying yellow pages of a mobile terminal. The method comprises: receiving a yellow page query request of a user, obtaining query information from the yellow page query request; according to a preset query condition, determining to query a local database or a server database through a network, the preset query condition including whether the mobile terminal is connected with a wireless fidelity Wifi or not, whether electric quantity of the mobile terminal exceeds a preset value or not, and whether time of the query information from previous query exceeds a preset time or not; and according to the query information, querying corresponding yellow page information in the local database or the server database, and returning a yellow page information query result to the user. The method and the device realize dynamic intelligent yellow page query.

The invention is applicable to the technical field of electric power big data application, and provides an abnormal water and electricity consumption detection method and device and terminal equipment. The method comprises the following steps: acquiring water consumption information and electricity consumption information of a plurality of users within a preset time, and determining water consumption and electricity consumption proportion information of each user according to the water consumption information and the electricity consumption information; Calculating a proportion information distance between the water and electricity proportion information, and establishing a proportion information distance matrix; And determining whether the user uses the hydropower station abnormally according to the proportional information distance matrix and the hydropower station proportional information of each user. Detection time can be saved, and the efficiency and accuracy of electricity stealing and water stealing detection are improved.

The invention provides a wireless sensor network target tracking method based on Fisher information distance. And after the target enters the monitoring area and is detected, the tracking system is activated to start tracking the target. In order to improve the tracking precision of the system, the use of more sensors to track a target is a conventional thought. However, the use of excessive sensors on the one hand consumes excessive energy and on the other hand is found by enemies due to the radiation of excessive energy so as to be locked in position. In order to solve the problems of radiation energy and tracking precision, on the basis that the number of sensors is not changed, a more efficient sensor is selected to track a target, and the tracking precision is improved. The method hastheoretical guidance significance and practical value for improving the concealment of a tracking system and prolonging the service life of the sensor network in some specific scenes.

The invention relates to the technical field of image communication, in particular to a peripheral resource pushing system and method for IPTV enterprise users. The system comprises a resource integration subsystem used for classifying information contents to generate an information classification table and generating an information score table according to information scores; an information pushing subsystem used for acquiring the current time, acquiring the information category according to the current time and acquiring the information content according to the information category and the information classification table; a server also used for obtaining a preset distance, generating an information distance according to the user positioning and the information content, and screening theinformation content according to the information distance and the preset distance; and an information pushing module also used for obtaining the information score from the information score table according to the screened information content, generating a pushing proportion of each information content according to the information score, and pushing the information according to the pushing proportion. By adopting the scheme, the information can be accurately pushed to the user according to the time and the position, and the user experience is improved.

The invention provides a multi-source heterogeneous data processing method and device and a storage medium, aims at a water surface and underwater space cross-domain communication system, realizes multi-source heterogeneous signal processing based on single equipment, and meets the cross-domain interaction requirement of water surface and underwater communication information. The multi-source heterogeneous signal processing device comprises a communication module, a control module and a signal processing module; the communication module is used for receiving different types of communication signals sent by external equipment; the control module is used for carrying out system task scheduling, state management and control, scene switching and storage forwarding related control and calling;the signal processing module is used for analyzing the received signals under the scheduling of the control module to obtain signal frequency spectrum characteristics and data characteristic information, and the signal frequency spectrum characteristics and the data characteristic information comprise at least one of azimuth information, distance information, energy information and moving speed information; and a target object is identified according to the extracted signal frequency spectrum characteristics, and the target object is tracked and positioned according to the extracted data characteristic information.

The present application provides a character correction method, device, device, and storage medium, wherein the method includes: the position information of the user's click operation on the terminal screen, and obtaining the feature information of the click operation; determining the position information and the position information on the terminal screen The distance information between each button; determine the transition probability information from the character indicated by the historical click operation to the character corresponding to each button; determine the character prediction result according to the feature information, distance information and transition probability information, and the character prediction result includes click The character indicated by the operation belongs to the possibility score of the character corresponding to each button; the character with the highest possibility score is used as the correction character corresponding to the click operation. Characters can be corrected based on the keyboard layout of the soft keyboard on the terminal, so as to avoid correction errors for different terminals, and correct the characters clicked by the user.

The invention relates to a data processing method and system based on a neural network, and relates to the field of intelligent transportation. The method comprises the steps that according to map data of a first area, an input feature vector of a neural network is constructed, and the map data comprises roadside information, distance information, road information and receiving information; constructing a training set, a verification set and a test set based on historical data; constructing a shallow neural network model; training, verifying and testing the shallow neural network model to generate a final neural network model; and integrating the neural network model to an automatic driving controller, and determining first information related to the map data. According to the data processing method and system based on the neural network, on the basis of multi-dimensional information such as access time, frequency, road characteristics and RSU distribution characteristics, data processing priority ranking is more reasonable, cache resources are effectively released, and the purpose of storing map data actually needing to be cached is achieved.

The invention discloses a method and a system for determining slab staggering of to-be-assembled segments, and the method comprises the steps: obtaining distance information which comprises the distances from a distance measurement device to two to-be-assembled segments; and determining whether slab staggering occurs to the two to-be-assembled segments or not and the slab staggering amount duringslab staggering at least according to the distance information. According to the method, automatic distance measurement is achieved through the distance measurement device, whether slab staggering occurs or not is automatically judged and the slab staggering amount is calculated according to the distance information, it is avoided that a caliper is adopted for manual positioning to judge whether slab staggering occurs or not, errors are large, efficiency is low, accurate positioning and assembling of the segments are achieved, manual positioning is not needed, and manpower resource waste is avoided.