333results about How to "Improve the quality of training" patented technology

The invention relates to an intelligent training system for a vehicle driver based on an actual vehicle. The intelligent training system comprises a vehicle-mounted intelligent training subsystem, a data processing subsystem, an operation management and control subsystem and a system-level data communication subsystem, wherein the vehicle-mounted intelligent training subsystem is mounted on the actual vehicle and is used for training interactive driving of the driver and acquiring and reporting data of a driver training process; the data processing subsystem manages basic information and is used for receiving, storing and processing the data of the training process acquired and reported by the vehicle-mounted intelligent training subsystem; the operation management and control subsystem monitors an operated state and a running state of the vehicle in real time and remotely controls the monitored vehicle according to actual conditions; the vehicle-mounted intelligent training subsystem is connected with the data processing subsystem and the operation management and control subsystem through the system-level data communication subsystem and realizes bidirectional information interaction.

The invention belongs to the field of aviation training, relating to a flight training data acquisition and quality evaluation system of an analog machine and a technical scheme thereof, in particular to the problems of acquiring the videos of flight data in a training system monitor of the analog machine, preprocessing an image after video frame graduation by applying an image processing technology, recognizing the flight data in the image by using a pattern recognition technology, carrying out three-dimensional simulation on a flight path in the training process of the analog machine so as to further carry out teaching evaluation, generalizing and summarizing typical problems, submitting the problems and flight training videos to a video server and providing a share and study platform. Under the conditions of keeping fully isolation with an analog machine system and avoiding a data communication channel with the analog machine system, the flight training data acquisition and quality evaluation system provided by the invention accurately acquires the flight data, reproduces the training process of a pilot and evaluates the quality.

Methods and apparatuses are provided for performing sequence to sequence (Seq2Seq) speech recognition training performed by at least one processor. The method includes acquiring a training set comprising a plurality of pairs of input data and target data corresponding to the input data, encoding the input data into a sequence of hidden states, performing a connectionist temporal classification (CTC) model training based on the sequence of hidden states, performing an attention model training based on the sequence of hidden states, and decoding the sequence of hidden states to generate target labels by independently performing the CTC model training and the attention model training.

The invention provides a motion-aid training system based on motion recognition. The system comprises a motion collection module and a motion-aid training module, wherein the motion collection module is used to acquire limb posture inertial data of to-be-trained motions of a trained user and send the data to the motion-aid training module; and the motion-aid training module conducts three-dimensional recovery demonstration based on the limb posture inertial data, and carries out motion assessment and motion bias correction training after human body motion recognition. According to the invention, based on a human body motion recognition method, the three-dimensional recovery demonstration, motion recognition and accurate evaluation are conducted on the training motions, so bias correction training can be carried out; and the user can objectively know personal motion errors and weaknesses and solve various problems encountered in learning training fundamentally, so that more targeted training and motion correction can be conducted and training effects can be improved.

The invention discloses a driver examination and training three-dimensional virtual monitoring method and system, and relates to the technical field of driver examination systems. The driver examination and training three-dimensional virtual monitoring method includes the steps that an examination site and the appearance of an examination vehicle are accurately measured through the satellite positioning technology, and measurement data are recorded; a three-dimensional graph of the examination site and the examination vehicle is generated according to the measurement data; a three-dimensional virtual scene is established according to the established graph of the examination site and the examination vehicle; a vehicle-mounted computer of the examination vehicle receives examination vehicle positioning coordinate data information acquired through satellite positioning, and the information is loaded to the three-dimensional virtual scene after being processed. By the adoption of the driver examination and training three-dimensional virtual monitoring method and system, the examination vehicle is monitored in an omni-bearing mode, positioning is accurate, and the established three-dimensional virtual scene is accurate and virtual and visible in an omni-bearing mode; monitoring personnel in an examination monitoring center can more accurately conduct item judgment on the examination vehicle and can find the misjudgment and judgment missing problems in examinations in time.

The invention discloses a student academic performance prediction method and a student academic performance prediction system based on a Naive Bayes model. The method comprises the steps of acquiring learning data of students, and transmitting the acquired data to a database of a first server; implementing data conversion on the learning data of the students stored in the database to acquire a standardized student learning condition data sheet; aiming at the standardized student learning condition data sheet, computing a conditional probability of each property in different classifications and prior probabilities of different classifications by a computation unit in the server; and implementing data conversion on the data of the student to be predicted and inputting the converted data to the trained Naive Bayes model to implement classification prediction on the student data. According to the method and system of the invention, the Naive Bayes model is adopted, and the model is used for accurately predicting the academic performance of the student, with high reliability.

The invention provides an (augmented reality) and VR (virtual reality) combined cornea smart operation training system. The system collects animal eyeball specimens during operation process and positive three-dimensional images of operation devices used by trainees in real time, performs the real-time side face 3D (three-dimension) scanning on the animal eyeball specimens, the operation devices used by trainees, and hand action of trainees through a TOF (time of flight) sensor, compares with preset data through a central processor and outputs guiding information and correction information, and sets up an augmented reality or virtual reality model containing the positive three-dimensional image, the guiding information and the correction information collected by a 3D microscope; through the AR glasses, the trainees can obtain the operation target object, status of the operation devices in use, the guiding information and correction information in real time, and thus the reality sense and on-site sense of cornea operation training are promoted, the real-time guidance in operation training is improved, the training operation of the cornea operation is shortened, the training cost of cornea is reduced, and the training quality is promoted.

The invention relates to a forklift operation training simulator. A main structure of the forklift operation training simulator comprises hardware equipment and a software system. The hardware equipment comprise a simulation cab, an steering control system, an instrument system, a multimedia computer, an audio system, an instructor control panel and the like. The software system comprises an operation platform and is used for real-time working environment animation generation of the computer, dynamic forklift driving simulation, sound simulation, operation evaluation, data management, network control and the like. The forklift operation training simulator has the advantages that a whole training period of a student can be shortened greatly, the student is enabled to finish training tasks in a short time, sudden incidents which rarely happen during practical training can be simulated conveniently, accurate-judging ability, good psychological quality and the ability of handling emergency circumstances correctly of the student can be cultivated, and training quality is improved.

The invention provides a power skill training platform based on a virtual reality technology. A virtual environment is programmed by 3D digital modeling of a power production scene through a graphic image generation and display unit set in a computer system; the virtual environment is displayed to a user equipped with 3D display equipment through the 3D display equipment of an interactive sensing system, and the line-of-sight direction of the user is measured by a set detection device so as to adjust the scene corresponding to the virtual environment; and the gestural movement of the user equipped with 3D space control equipment is detected through the 3D space control equipment, and is fed back into the 3D display equipment to adjust the scene corresponding to the virtual environment. By building an immersive virtual environment based on high-risk, invisible and unreproducible scenes in power production, the quality of training is improved effectively, the training efficiency is improved, and the training cost is reduced.

The embodiment of the invention provides a ball serving control method and a mobile terminal. The ball serving control method comprises the steps of obtaining multiple frames of images through a camera; identifying the main body in each frame of images of the multiple frames of images, wherein each main body comprises at least one of training personnel, a racket and a ball; obtaining motion information of the training personnel according to the position of each main body in the corresponding frame of images and the time interval of every two adjacent frames of images, wherein the motion information of the training personnel comprises at least one of the moving direction, the moving speed, the ball serving direction and the ball serving intensity of the training personnel; determining the ball serving information according to the motion information of the training personnel; controlling ball serving equipment to adjust the ball serving mode based on the ball serving information. According to the embodiment of the invention, the ball serving mode can be dynamically adjusted according to actual situations of the training personnel, so that training is more targeted, and then the training quality and the training effect are effectively improved.

The invention relates to an optimized-neural-network gesture-recognition human-computer interaction method based on GL. Gestures of an operator are collected through a Data Glove14 Ultra data glove are collected, then the collected data is processed, an optimized neural network based on the GL algorithm is subjected to network learning training, finally, the trained algorithm is applied to gesture recognition based on the data glove, and teleoperation of the robot is achieved.

The invention relates to a small sample target detection method based on attention and contrast learning, and belongs to the field of artificial intelligence and image processing. The invention relates to the small sample target detection method which combines data enhancement, an attention region suggestion network (Attention RPN) and comparative learning. The method is based on a Faster R-CNN (Convolutional Neural Network) network, and comprises the following steps of: adopting a Few-shot Mosaic data enhancement module for enriching the comparison between a small sample background instance and a Novelclass instance and a Base class instance, enhancing the attention of a model on a foreground by an Attention RPN (Regression Coordinate Compensation) module based on regression coordinate compensation, and improving the expression of instance-level features by a contrast learning module. According to the method, the new class detection precision of the Faster R-CNN on a small sample is improved, and meanwhile, relatively high base class detection precision is kept; the dependency of Faster R-CNN on the new class training sample size is reduced, the new class migration ability is improved, and the effectiveness of the method is verified on COCO and VOC data sets.

The invention relates to a percussion training system based on motion capture. The system comprises a percussion simulation object, a digital 3D model processing system and an optical motion capture system. The digital 3D model processing system comprises a digital 3D modeling unit, a model case matching unit and a database. The optical motion capture system comprises an acquisition unit and a digital 3D calculating processing unit. In the invention, a simple percussion simulation object is converted into a digital object matching with the size of the percussion simulation object, and a case and a sound associated with a percussion position of the digital object are established. An optical motion capture technology is combined to carry out spatial three-dimensional positioning of a percussion finger, real percussion sounds of the percussion position under different operation methods are simulated and realized, the digital object in a digital scene can be directly observed so that an operator can conveniently realize an anatomical structure corresponding to the percussion position during a practice process; a diagnosis ability is learnt; and percussion training quality and an effect are increased.

The invention provides an ophthalmologic operation training system and method based on movement capturing and mixed reality. According to the ophthalmologic operation training system based on movementcapturing and the mixed reality, front three-dimensional images of an animal eyeball specimen during an operation are collected in real time through a 3D camera microscope, an operation tool is positioned by arranging a pulse laser detector on the operation tool, the position of the operation tool is compared with a standard position through a central processor, position correcting information isgiven, guide information is given through the central processor according to operation processes, then, the front three-dimensional images collected by the 3D camera microscope are superposed with the position correcting information and the guide information, and display is carried out through a mixed reality display terminal. By the adoption of the ophthalmologic operation training system and method based on movement capturing and the mixed reality, the senses of reality and immediacy of ophthalmologic operation training can be improved, real-time guide effects in operation training are improved, ophthalmologic operation training periods are shortened, ophthalmologic operation training costs are reduced, and the training quality is improved.

The invention discloses a precaution analysis system for a distribution line and an analysis method. The analysis system and analysis method are characterized by comprising gloves, a vibrator, a somatosensory device, a host machine and a display screen; the host machine is respectively connected with the gloves, the vibrator, the somatosensory device, and the display screen; the gloves are data gloves and internally provided with a temperature sensor and a humidity sensor which are used for sensing hand temperature and humidity data; the vibrator generates an impulse vibrating signal when wrong operation is happened, stimulates trainees and enhances the training effect; the somatosensory device catches movement action of human body, and displays corresponding operation action of a virtual person on the display screen. The precaution analysis system can realize interactively, lively and vividly staff safety training, and improve the electrical safety production knowledge and skill training quality.

The invention provides a basketball pass training device to solve the problem in the existing basketball pass training device that the passing angle is single, the passing quality is unstable, basketballs are not launched continuously, and basketball launching is limited by the shooting angle of trainees. The training device does not require personnel to carry out docking pass training to save manpower, and a cylinder is adopted to launch the basketballs, so that the passing quality becomes stable and reliable. At the same time, by means of alternative engagement of a third sector gear and a fourth sector gear with an arc rack, the angle of the launching cylinder in the vertical direction is adjustable. A horizontal angle adjusting device is also added to enable a swing rod to swing with an eighth gear to achieve angle adjustment of the launching cylinder in the horizontal direction. The basketball launching angle is varied, and the basketball catching difficulty of the trainees is improved. At the same time, the basketballs can be launched continuously when the trainees miss the basketballs, and the basketballs can be launched quickly when the trainees make a goal. The training time is saved, and the training quality is improved.

The invention provides a method for identifying an air control area where an airplane is located. The method is characterized by comprising acquiring flight parameters of a simulation machine in real time; extracting longitude, latitude and altitude data of the airplane from an acquired flight parameter set of the simulation machine; traversing all air control area data in an air control area data set by the aid of a specific algorithm and altitude comparison; acquiring zero to a plurality of air control areas with matched longitude, latitude and altitude data by means of comparison matching. The method has the advantages that a simulation machine system which implements the method can quickly identify and determine an area such as the control area, a secondary sector of a local control area, an approach control area, a tower control area and another air control area where the airplane is located according to the flight parameters of the simulation machine, requirements of the analog simulation machine system are met, the simulation machine can replace a real civil aircraft to realize a flight training effect in normal subjects and special subjects, the training quality is effectively improved, and the training cost is effectively reduced.

The invention belongs to the technical field of image processing, and particularly relates to a target detection method and device for a parallel robot vision system, and the method comprises the steps: collecting a target image, and dividing the target image into a training data set and a test data set; building a hybrid automatic encoder, feeding images in the training data set into the hybrid automatic encoder in batches for denoising and sparse processing, and obtaining image data with classification labels; Sending the image data into a Faster RCNN neural network to carry out model training; Sending the test data set image data into a trained Faster RCNN neural network to carry out model test; And embedding the tested Faster RCNN neural network model into a parallel robot vision system, and carrying out target identification and positioning on the to-be-detected image. According to the method, the model anti-interference capability and robustness are improved, the data set training difficulty is reduced, the model training speed and quality are improved, the neural network is easy to reuse for different to-be-identified targets, and the method has important guiding significance for the technical field of image processing and target identification.

The invention discloses a sample extraction method and device for a video classification problem, and relates to the technical field of video classification. the method for extracting the samples fromthe video data is optimized so as to ensure that the samples can cover the whole content information of the video data, meanwhile, redundant image data in the samples are avoided, the sample qualityis enhanced, and the training quality of a video classification model can be improved. The technical scheme includes; acquiring video data; Analyzing the video data to obtain a plurality of continuoussingle-frame images corresponding to the video data; extracting Feature images from the plurality of continuous single-frame images to form a sample, wherein the feature images are used for summarizing content information of the video data, and the sample does not contain redundant image data. The method is applied to extracting the samples for training the video classification model from the video data.