152results about How to "Real-time output" patented technology

The invention discloses an inertia measuring unit, which comprises a measuring short section frame, a first dynamically tuned gyroscope, a rotary mechanism frame, a second dynamically tuned gyroscope, a first accelerometer, a second accelerometer, a third accelerometer and a motor; the first dynamically tuned gyroscope is arranged on the measuring short section frame along the radial direction of the measuring short section frame; the second dynamically tuned gyroscope is arranged on the rotary mechanism frame along the axial direction of the rotary mechanism frame; the first accelerometer, the second accelerometer and the third accelerometer are orthogonally arranged on the axial direction of the rotary mechanism frame along the axial direction of the rotary mechanism frame in turn; the second dynamically tuned gyroscope is arranged between the first dynamically tuned gyroscope and the first accelerometer; the measuring short section frame and the rotary mechanism frame are coaxial; and the motor drives the rotary mechanism frame to rotate around the axis of the rotary mechanism frame. The inertia measuring unit and the continuous inclinometer for the dynamically tuned gyroscope of the invention realize all-round measurement and continuous measurement of well tracks and improve the measurement efficiency.

The invention relates to a video integration system, and provides a communication protocol of a computer serial port control console and a novel infrared and visible video integration processing system. The novel infrared and visible video integration processing system comprises a console control portion, a camera video acquisition portion, a video integration portion and an integration displaying portion. A vc++ coding controlling and displaying interface is utilized, and the opencv (open source computer vision library) is introduced to process video images, and then an integration result is displayed. Console control includes the control of upward rotation, downward rotation, leftward rotation, right rotation and rotating speed, a thermal infrared imager and a visible light camera are utilized in the camera video acquisition portion to acquire two video information, and zoom and the like of the camera can be controlled via a control interface. By means of a novel visible light and infrared image registration method, the infrared and visible video integration system is simple and quick, and instantaneity of the infrared and visible video integration system which is simple and quick is strengthened. Since the infrared image and visible light image can greatly compensate mutually, the infrared and visible video integration system is high in practicability, good in integration effect and wide in application.

The invention relates to a horizontal well logging technology with a continuous tube and a device thereof. The technology uses a high-strength and high-tenacity continuous tube as tools to convey down-hole logging instruments so as to ensure that the down-hole instruments perform tests under various horizontal wells. The adopted device comprises a continuous tubing unit, the continuous tube, a wellhead device, connecting joints between the continuous tubes and the down-hole instruments, down-hole logging instruments and tools for the horizontal wellbore, a ground numerical control logging system, a depth system of the continuous tube and a signal transmission system of the continuous tube. The invention can ensure the down-hole instruments to perform tests under various horizontal holes by using the continuous tube as the conveyance tool of the down-hole instruments; the technology uses few application equipment and has the advantages of good maneuverability, convenient operation and high construction efficiency thanks to the application of continuous tubing unit and other special devices.

The invention discloses a radar detection system and method for low-altitude multi-target classification and identification. The radar detection system for low-altitude multi-target classification andidentification includes a waveform generator, an antenna unit, a reception unit and a signal processing unit, wherein the waveform generator is used for generating detection wave beams in which the energy is concentrated below the specified threshold according to the low-altitude area to be detected; the antenna unit is used for transmitting the detection wave beams output from the waveform generator and receiving a target echo signal; the reception unit is used for performing down-conversion on the target echo signal received by the antenna unit to become an intermediate frequency signal andoutput the intermediate frequency signal; and the signal processing unit is used for acquiring the positional information of the target according to the intermediate frequency signal output from thereception unit and outputting the identified target type. The radar detection method for low-altitude multi-target classification and identification is to realize low-altitude multi-target classification and identification. The radar detection system and method low-altitude multi-target classification and identification can realize real-time detection and classification and identification of low-altitude multiple targets, and have the advantages of high identification efficiency, high accuracy and high flexibility.

The invention provides a method for projected display on a textured plane in the technical field of computer-based image processing. In the method, firstly, a radiation model is established for showing that after the light output by a projector is reflected by a projection screen, the light is grabbed by a camera; secondly, the luminosity response function of the projector and the camera, and the color mixing matrices of the projector and the camera are measured; thirdly, the geometrical mapping between the projector and the plane of the camera are established through projected checkerboard images; fourthly, with the given luminosity response function of the projector and the camera, and the color mixing matrices of the projector and the camera, two gray scale images are projected, and the reflectivity information of the projected surface and the ambient light information are found; fifthly, the projected images are rectified and are projected again before the projected images are output according to the geometrical mapping as stated in the third step and the reflectivity information of the projected surface and the ambient light information as stated in the fourth step. After the projected images are rectified and the light is accelerated through a GPU, the speed can reach 60fps and meets the requirement for real time. The method is applicable to various practical circumstances.

The invention discloses a device and method used for measuring axial bolt relaxation in consideration of joint surface characteristics, which belongs to the field of mechanical manufacturing and structural mechanics. A bolt bar passes through upper and lower connectors, and contact surfaces have specific size and surface data. A pressure sensor is used to measure the attenuation of the bolt clamping force. An eddy current displacement sensor is used to measure displacement between the contact surfaces. Three-dimensional topography scanning is used to measure the surface three-dimensional data. A tensile testing machine is used to carry out cyclic tensile test and simulate the bolt conditions. The axial clamping force and axial displacement are transmitted to a computer for recording and processing through a data acquisition system. Axial bolt relaxation characteristics are acquired by analysis and processing several kinds of data. According to the invention, the connection characteristic of an axial part is ensured; the influence of a fixed joint surface on bolt relaxation is considered; and the bolt relaxation test precision is improved.

The invention discloses a distributed coal-rock dynamic disaster potential real-time monitoring method and system. A plurality of monitoring electrodes are distributed on the surface of a coal-rock gangway or distributed into a drill hole. A potential monitor continuously collects potential signal differences in real time, wherein the potential signal differences are generated by loads of a coal-rock mass. According to distribution and changes of the potential signal differences, the mode that area critical values and dynamic tendency are combined is adopted to conduct real-time early warning on coal-rock dynamic disaster risks and risk areas. According to the distributed coal-rock dynamic disaster potential real-time monitoring method and system, the loads of the coal-rock mass, the deformation breakage degree, the dynamic disaster risk distribution and evolution situations are reflected through distributed monitored potential signals, and the distributed coal-rock dynamic disaster potential real-time monitoring method and system have the advantages of being good in real-time performance and responsiveness, high in anti-jamming capacity and automation degree, higher in practicability and convenient to install.

The invention discloses an image registration method and an image registration device. The image registration method includes acquiring a second group of feature points of thermal imaging images; subjecting the second group of feature points to registering change to acquire a third group of feature points, wherein the registering change is based on registration parameters acquired in advance, an image acquiring device of visible images, under each focal length, has a registration parameter corresponding to the focal length, and the corresponding registration parameter under each focal length is determined according to the linear relation between the registration parameters and the focal lengths; determining a fourth group of feature points, same as feature points of the visible images in position, in the third group of the feature points; after determining the registration parameters according to the fourth group of feature points, registering the visible images and the thermal imagingimages according to the registration parameters. By the image registration method and the image registration device, real-time registration of the invisible images and the thermal imaging images canbe realized during focal length change of a visible light camera, and the visible images and thermal-imaging blended images can be output in real time.

The invention discloses a method and device for adding punctuation in voice recognition, a computer device and a storage medium. The method comprises: performing voice recognition on an acquired to-be-recognized voice, synchronously detecting a mute segment in the to-be-recognized voice, and determining whether the time length of the mute segment exceeds a first time length; if so, outputting a text sequence before the mute segment, and inserting a comma or period into a corresponding position in the text sequence according to the time length of the mute segment; performing voice recognition on the acquired to-be-recognized voice after the mute segment, and correcting the comma or period inserted into the text sequence according to a preset discrimination model. The method and device are used for improving the output efficiency and the accuracy of the punctuation in voice recognition, so as to achieve a purpose of improving voice recognition efficiency, accurately breaking sentences and accurately expressing emotions.

The invention provides an abnormal behavior detection method and system based on a human skeleton, and a medium, and the method comprises the steps: S1, building a training video sample database, andbuilding an original video database with a preset frame length according to a public violent behavior database; s2, according to the obtained original video database, extracting human body skeletons in all video samples in the original database by utilizing an open-source attitude estimation algorithm OpenPose to obtain a skeleton video database; s3, deframing all the obtained video samples in theskeleton video database into a skeleton picture stream by using an FFmpeg tool to obtain a skeleton picture stream; S4, sequentially inputting the obtained skeleton picture streams into a 3D-DenseNetnetwork model for training, so as to obtain a trained 3D-DenseNet network model. The human skeleton and the 3D-DenseNet are used for abnormal behavior detection for the first time, the detection accuracy is high, the robustness is high, and the detection result can be output and updated in real time.

The invention relates to a ship target situation simulation method based on electronic chart, and the technical feature of the method is that it comprises the following steps: (1) establishing multiple simulation target location information based on the electronic chart information; (2) Setting initial motion parameters of each simulation target respectively; (3) Establishing a planned route for each simulation target; (4) Performing risk parameter calculation for multiple planned routes; (5) Repeating step (3) and (4) if the calculated risk parameter reaches the threshold in the set safe navigation area; (6) Superimposing the annotation information on the electronic chart information; (7) Calculating the dynamic target information in real time and outputting it to the business applicationplatform with corresponding requirements in real time; (8) Establishing a target situation simulation plan library. According to the ship target situation simulation method based on electronic chart,the simulation situation targets are generated on the basis of electronic chart information, and target motion situation drillings are carried out in real time according to different parameter setting and routes planning.

The utility model belongs to an image shooting field and provides a multi-scene image pick-up device, comprising a plurality of image sensing modules aiming at different angles and simultaneous image shooting, a complex image processing module for composing a plurality of images collected by various image sensing modules in real time and sending the images out, and an integral sequential control module for generating sequence signals required by the image sensing modules and the complex image processing module. The image pick-up device in the utility model can shoot images simultaneously with a plurality of image sensing modules aiming at scenes in different angles; carry out sequential control with the complex sequential control module, and then compose the images in different angles through the complex image processing module before real-time output. Therefore, the image shooting device has advantages of simple circuit structure and low cost. Moreover, the image shooting device can simultaneously shoot images at different angles, hence wider field range of vision and real-time output effect.

The invention discloses a high-precision target positioning and speed measurement method based on a photoelectric platform of an unmanned aerial vehicle. The high-precision target positioning and speed measurement method provided by the invention includes the following steps: first, an azimuth error and a high-low angle error of the photoelectric platform are estimated according to a stationary cooperative target, after estimated values of the azimuth error and the high-low angle error of the photoelectric platform are stabilized, the stabilized estimated values of the azimuth error and high-low angle error of the photoelectric platform are utilized as the azimuth error and the high-low angle error of the photoelectric platform, and then compensation is performed on the azimuth error and the high-low angle error of the photoelectric platform in real time during a positioning and speed measurement process, and a high precision target position and speed information are output.

The invention discloses a bird population recognition and analysis method based on a GoogLeNet network model. A training picture sample database is established; a GoogLeNet network a capable of judging bird pictures is obtained by using picture samples to train the GoogLeNet network model; a GoogLeNet network b capable of accurately judging bird populations is obtained by using picture pictures totrain the GoogLeNet network model; frames of to-be-identified videos input in real time are decoded into to-be-identified picture flows; each frame of pictures in the picture flows is sequentially input into the GoogLeNet network a, and whether the pictures are bird pictures or not is judged; if yes, the pictures are input into the GoogLeNet network b, the included bird populations are identified; picture identification result flows are obtained, and final identification results are output from the picture identification result flows. The bird population recognition and analysis method fillsthe blank of bird population identification based on a deep learning model, the identification accuracy is high, and the identification results can be output and updated in real time.

The embodiment of the invention provides a video synthesis method and system. The method is applied to a guide station of the video synthesis system. The method comprises the following steps: receiving a video synthesis instruction, wherein the video synthesis instruction at least comprises a video main body and a video picture from different video sources; determining a first video source where the video main body is located and a second video source where the video picture is located; extracting target video data of the video main body in the first video source according to a preset algorithm; and superimposing the target video data to a preset location of the second video source. By adoption of the video synthesis method and system provided by the embodiment of the invention, on the basis of the existing live broadcast / recorded broadcast system, image matting and synthesis are realized through the guide station, the operation is convenient, and the original high-definition video data are retained; and for a live video source, the real-time output can be performed after the synthesis, thereby shortening the delay and meeting the live broadcast use demands.

The invention discloses a course updating method in a vehicle navigation system, which comprises the following steps: firstly, respectively setting two time thresholds and a course difference threshold; and updating the course of a vehicle according to a relation between a difference value of course difference and GPS directly input course difference, and a relation between gyro continuous integral time and a time threshold, which are obtained according to different running states of vehicles, that whether the course spans 360 DEG, and gyro data integral of five recently continuous points. The method is simple and easy to realize, can obviously improve the positioning accuracy of the navigation system, can update course output in real time, and improves the real-time, reliability and accuracy of the navigation system.

The invention provides a residual moment measuring device and method of an inertial device based on three-dimensional air flotation, belonging to the technical field of detection. An on-stage system and an under-stage system in the device are connected through a wireless transmission system; the under-stage system comprises an under-stage data collecting and processing system, a server, a cabinetbox and an air supply system; the on-stage system comprises a power supply system, an inertial execution mechanism control system, an inertial execution mechanism to-be-tested product, an on-stage data collecting and processing system, a residual moment test system, a vibration isolating and supporting system, and a vacuum control system, the residual moment test system consists of an air flotation rotary table, a high-precision sensor device and an anti-overturning and protective device, and the high-precision sensor device is separately arranged in a vertical direction of the air flotation rotary table and in a horizontal direction outside the air flotation rotary table. The device can simulate orbit work of a satellite platform, provides a simulated spatial mechanical environment, outputs residual moment in real time, can be used for directly measuring multiple inertial devices, is low in calculation amount, and has high precision.

This invention refers to angle reference laser digital sensor, which contains laser set on casing top, casing, optical facula modulation system, liquid double optical wedge compensator, PSD position photoelectric sensor and signal output line connected to outside of casing, the principle of said invention is that the laser beam specially modulated by optical system automatically generates high precision vertical laser beam, the standard deviation reaches 1 / 600000, the compensation precision greatly exceeds the traditional technology with real time outputted digital signal.

The invention discloses an outdoor multimedia full-automatic floor sweeping robot, including an outdoor navigation system, a cleaning system and a multimedia system. the outdoor navigation system, thecleaning system and the multimedia system are controlled by the master controller, the outdoor navigation system includes a drop prevention sensor, Video and Sound Monitoring Module, RTK GNSS receivemodule, Wireless communication modules and lidar, drop sensor, video and sound monitoring module, The RTK GNSS receiving module, the wireless communication module and the laser radar are all connected on the master controller. The invention realizes the geographic position coordinates of the cleaning robot point through the independently developed RTK and the laser radar data fusion algorithm andcomplementary filtering, and accurately constructs the precise three-dimensional scene map, which has stronger robustness.

The present invention discloses a method and a device for measuring a visual pose of a simulated satellite on an air floating platform. The method includes the following steps: arranging a cooperativetarget, establishing a world coordinate system to obtain three-dimensional coordinates of the cooperative target; obtaining a current frame image in real time by using a camera module; processing theimage to identify the cooperative target, matching two-dimensional pixel coordinates with the three-dimensional coordinates to obtain a matching point pair; and obtaining the pose for the matching point pair by solving according to an PNP algorithm. The present invention has low cost and is simple to be implemented, and it can be ensured that the cooperation target is included a camera vision ina moving range of the analog satellite, so that the pose can be stably outputted in real-time.

The invention discloses an image adaptive synthesis method and device and a computer readable storage medium, and belongs to the technical field of image processing. The image adaptive synthesis method comprises the following steps: firstly, determining camera parameters of a foreground and a background; then, respectively obtaining any matching point corresponding to the synthetic track in the foreground and the background, and determining coordinates of the two matching points in a camera coordinate system and coordinates of the two matching points in a pixel coordinate system; and then, judging the depths of the two matching points in a camera coordinate system, carrying out affine transformation processing on the foreground or the background, and carrying out image synthesis through OpenCV to obtain a synthesized image. According to the invention, affine transformation processing is firstly carried out on the foreground or the background, and then image synthesis is carried out through OpenCV, so that the purposes that control parameters do not need to be manually adjusted in the shooting process and the synthesized image can be output in real time can be achieved under the condition that the positions and parameters of the corresponding foreground camera and background camera are different.

The invention discloses a wave height recorder and accelerometer double-purpose calibration device. The device comprises an intelligent combined unit 1, a working table 2, a wave height reorder 3, a rotary bracket 4, a rotary spindle 5, an accelerator 6, a control box 7, an adjusting screw 8, a frame 9, a servo motor 10, an encoder 11 and a water cylinder 12, wherein the working table is in a vertical state when the rotary bracket 4 is rotated to a vertical position to be mounted by means of the rotary spindle 5, the wave height recorder mounted on the working table does reciprocating motion up and down, and the lower part of the working table is provided with the water cylinder for simulating seawater level so as to calibrate the wave height recorder; the working table is in a horizontal state when the rotary bracket 4 is rotated to a horizontal position to be mounted by means of the rotary spindle 5, and the accelerometer does horizontal reciprocating motion to calibrate the accelerometer; a control system controls a rotation speed and reciprocating convolution of a servo motor photoelectric encoder of the intelligent combined unit and controls the frequency and the amplitude of the calibration device; and compared with corresponding values of the wave height recorder and the accelerometer, the value achieves higher calibration accuracy.

The invention discloses a concrete construction nondestructive testing method based on a geomagnetic pulsation technology. The concrete construction nondestructive testing method is characterized by comprising the following steps of: (1) arranging a pulsation signal detecting and processing system; (2) mounting a three-way pulsation sensor at one part with a good lining integrity; transmitting a collected pulsation signal to a data analysis processor by a pulsation signal demodulator; and using the received data as a reference data by using the data analysis processor; and (3) mounting a three-way pulsation senor at one part needing to be detected; transmitting a collected pulsation data to a data analysis processor by a pulsation signal demodulator; comparing and analyzing the received data and the reference data by using the data analysis processor; and judging the lining integrity according to a result. The concrete construction nondestructive testing method disclosed by the invention has the advantages of high accuracy of the collected signal, capability of collecting the three-way pulsation signal, no need of providing seismic source by using a natural oscillation pulsation signal of the lining as well as capability of being widely applied to health detection of concrete constructions, such as tunnels, water and electricity, slope protection and the like.

The invention provides a video real-time processing and synthesizing system and method. The system comprises a video acquisition device, an audio acquisition device, a video acquisition height adjustment device, a display adjustment device, a video matting synthesis device and a real-time output device. According to the invention, the switching of a plurality of visual angles and real-time combination of multiple backgrounds are achieved, automatic adjustment and correction can be carried out according to the height of a portrait, the optical axis of the camera and the optical axis of a humaneye are automatically adjusted to be unified, processed videos can be provided for a user in real time, used equipment is simple and convenient, cost is low, operation technical difficulty is low, andthe requirements of common people can be met.

The embodiment of the invention provides a moving target detection method, detection device, and a mobile detection system, so that a problem that target echo signals that are detected by the existing mobile detection system during moving target detection include echo signals of fixed targets and thus lots of false alarm targets are generated in the prior art is solved. The method comprises: while a mobile carrier moves, attitude measurement of auxiliary equipment as well as moving speed compensation information obtained and transmitted by a navigation system is received; on the basis of the moving speed compensation information, a corresponding Doppler frequency deviation value of movement, relative to a target, of the mobile carrier is obtained; and according to the Doppler frequency deviation value of movement, relative to the target, of the mobile carrier, a detected target echo signal is compensated, an echo signal of a fixed target in the target echo signal is suppressed, and a Doppler frequency deviation value of the moving target is obtained, thereby obtaining accurate moving target detection distance information and radial speed information.

The invention relates to a precision motion control system and a control method thereof. The system comprises a control core module, a sensor input module, a high-voltage driving output module, a multichannel signal acquisition module, an executor module, a controlled object and a host computer screen display module. The output end of the sensor input module is connected with the input end of the control core module through the multichannel data acquisition module; the output end of the control core module is connected with the input end of the high-voltage driving output module; the host computer screen display module is in bidirectional connection with the control core module; the output end of the high-voltage driving output module is connected with the input end of the executor module; the output end of the executor module is connected with the input end of the controlled object and drives the controlled object; and the output end of the controlled object is connected with the input end of the sensor input module.

The invention discloses a deep learning method and system based on text implication identification of law and regulation and judicial explanation accurate matching and contradiction detection. A lawsand regulations or judicial explanation text is input, and the associated laws and regulations or judicial explanation text and the laws and regulations or judicial explanation text having content conflicts with the laws and regulations or judicial explanation text can be accurately matched in the data. By utilizing a text implication identification technology and a deep learning network, the input laws and regulations or judicial interpretation texts and laws and regulations or judicial interpretation texts in a database are subjected to accurate matching and contradiction detection, and a plurality of most relevant laws and regulations or judicial interpretation texts or a plurality of most relevant content conflicting laws and regulations or judicial interpretation texts are output. Theblank in the aspects of law and regulation and judicial interpretation text automatic matching tasks and law and regulation and judicial interpretation text conflict content detection tasks based ona deep learning network is filled.

The invention relates to a power grid real-time adaptive decision-making method based on deep reinforcement learning. The method comprises the following steps: step 1, modeling a novel power system unit adaptive scheduling problem as a Markov decision-making process; 2, researching the basic principle of an SAC algorithm, and solving a strategy enabling the accumulated reward value of the MDP model in the step 1 to be maximum; step 3, designing a neural network pre-training scheme based on behavior cloning in the IL, simulating expert experience, optimizing an original action space, proposing an IL-SAC algorithm, training a corresponding power grid optimization scheduling agent based on the IL-SAC algorithm and 105 real power grid scene data, and performing optimization scheduling on the power grid. During testing, the intelligent agent can output a real-time decision scheme in response to different power grid scene data, and intelligent regulation and control of a novel power grid system are realized. The power grid dispatching strategy can be output in real time.

The invention relates to a decision-making behavior generation method and system based on a decision-making tree, and the method comprises the steps: S1, carrying out the feature extraction from customer information, and obtaining an available index for the application of the decision-making tree; Ss2, forming a father / child relationship among all the single nodes according to the types, node codes and input / output parameters of all the single nodes in the decision tree; Ss3, performing condition configuration on each node to obtain a decision path, and performing verification to obtain a decision tree; Ss4, performing batch pre-running on the decision tree to obtain a decision tree batch pre-running result, and verifying the batch pre-running result to obtain an available decision tree; Ss5, configuring the batch running starting point of the available decision tree to obtain a decision tree batch running task and a decision tree version; and S6, performing batch running tasks throughthe decision tree of the system application to obtain decision behaviors which can be executed immediately. According to the method, tThe decision tree can be efficiently constructed, dynamic real-time positioning of clients can be realized, decision actions are timely output, and timely execution of client decision behaviors is effectively realized.

The invention provides a vehicle real-time tracking method and a computer readable storage medium, and the method comprises the steps: obtaining an image of each camera in a plurality of preset cameras, and obtaining a license plate region and license plate number information of a license plate of each vehicle in the image; performing coordinate conversion on the license plate region to obtain a first position of the vehicle in a world coordinate system; obtaining a second position of the vehicle in the real world through cooperative positioning of the plurality of cameras; and performing multi-camera cooperative tracking on the vehicle in real time based on the second position. By performing multi-camera cooperative positioning on the originally independent cameras, a more accurate vehicle position is obtained, so that multi-camera cooperative tracking is realized, and vehicle information is acquired in real time.