Abstract: A method and device for timing alignment of audio signals. The method includes: generating frequency domain images respectively for an audio signal to be aligned and a template audio signal (S110); inputting the frequency domain images into a twin neural network of a timing offset prediction model respectively, to obtain two frequency domain features output by the twin neural network (S120); fusing the two frequency domain features to obtain a fused feature (S130); inputting the fused features into a prediction network of the timing offset prediction model to obtain a timing offset output by the prediction network (S140); and performing timing alignment processing on the audio signal to be aligned according to the timing offset (S150). The technical solution is more robust, and especially in a noisy environment, features extracted by a deep neural network are more intrinsic and more stable. An end-to-end timing offset prediction model is more accurate and faster.

Abstract: The subject matter provides a path planning method, apparatus and electronic device. Wherein, the method comprises: performing environment modeling according to static road network information and dynamic road condition information of a road network so as to obtain an environment model; determining a plurality of candidate paths according to a starting point and an ending point; extracting from the environment model an environmental feature corresponding to each candidate path by a feature extraction network of a path planning model; inputting the environmental feature to a value estimation network of the path planning model so as to obtain an estimated value for each candidate path output by the value estimation network; determining an optimal path among the candidate paths according to the estimated value.

Abstract: A global path planning method and device for an unmanned vehicle are disclosed. The method comprises: establishing an object model through a reinforcement learning method, wherein the object model includes: a state of the unmanned vehicle, an environmental state described by a map picture, and an evaluation index of a path planning result; building a deep reinforcement learning neural network based on the object model established, to obtain a stable neural network model; inputting the map picture of the environment state and the state of the unmanned vehicle into the deep reinforcement learning neural network after trained, and generating a motion path of the unmanned vehicle. According to the present disclosure, the environment information in the scene is marked through the map picture, and the map features are extracted through the deep neural network, thereby simplifying the modeling process of the map scene.

Abstract: A multi-screen display system and a mouse switching control method are disclosed. The mouse switching control method is applied to a multi-screen display system comprising a main display screen and at least one extended display screen, and comprises: obtaining user images collected by cameras installed on the main display screen and the extended display screen respectively; inputting the user images into a neural network model, and predicting a screen that a user is currently paying attention to using the neural network model to obtain a prediction result; and controlling to switch a mouse to the screen that a user is currently paying attention to according to the prediction result. The system and mouse switching control method are based on self-learning of visual attention, predict the current screen operated by the user, automatically switch the mouse to the corresponding screen position, and improve the user experience.

Abstract: The present disclosure discloses a path planning method and device and a mobile device. The method comprises: collecting environmental information in a viewing angle by a sensor of a mobile device, processing the environmental information by using an SLAM algorithm, and constructing a grid map; dividing the grid map to obtain a plurality of pixel blocks, using an area constituted of pixel blocks not occupied by obstacles as a search area for path planning, and obtaining a processed grid map; determining reference points by using pixel points in the search area, and deploying topological points on the processed grid map according to the reference point determined and constructing a topological map; and calculating an optimal path from a starting point to a preset target point by using a predetermined algorithm according to the topological map constructed. The present disclosure improves path planning efficiency and saves storage resources.

Abstract: A method and device for extracting key frames in simultaneous localization and mapping and a smart device. The method includes acquiring an image frame from an image library storing a plurality of image frames of an unknown environment, and performing feature extraction on the image frame to obtain information of feature points, wherein the information includes a quantity of feature points; acquiring relative motion information of the image frame relative to the previous key frame, and calculating an adaptive threshold currently used by using the relative motion information; and selecting a key frame according to the information of feature points and the adaptive threshold indicating space information of image frames.

Abstract: A multi-AGV motion planning method, device and system are disclosed. The method of the present disclosure comprises: establishing an object model through reinforcement learning; building a neural network model based on the object model, performing environment settings including AGV group deployment, and using the object model of the AGV in a set environment to train the neural network model until a stable neural network model is obtained; setting an action constraint rule; and after the motion planning is started, inputting the state of current AGV, states of other AGVs and permitted actions in a current environment into the neural network model after trained, obtaining the evaluation indexes of a motion planning result output by the neural network model, obtaining an action to be executed of the current AGV according to the evaluation indexes, and performing validity judgment on the action to be executed using the action constraint rule.

Abstract: The present application discloses a method and device for updating a map. The method for updating a map according to the present embodiment includes: in a process of movement of a robot, when it is detected that an actual environment is different from an environment that is indicated by a global map that has already been established, starting up map updating, and establishing an initial local map; determining a locating point according to acquired sensor data and the global map, and optimizing the initial local map according to the locating point, to obtain an optimized local map; and covering a corresponding area of the global map by using the optimized local map, to complete updating of the global map. The embodiments of the present application improve the locating accuracy, ensure the speed and efficiency of the map updating, and save time.

Abstract: The present application discloses a method for filtering image feature points and a terminal. The method for filtering image feature points includes: providing quality score values to feature points extracted from an image, and according to the feature points and the quality score values of the feature points, training a neural-network model; after one time of filtering has started up, acquiring one frame of an original image and extracting feature points in the original image; inputting the original image and the feature points in the original image into the neural-network model, obtaining and outputting quality score values corresponding to the feature points in the original image; and according to the quality score values, filtering the feature points in the original image. The method for filtering image feature points can improve the success rate of the matching of the feature points in relocated application scenes, thereby improving the locating efficiency.

Abstract: A global path planning method and device for an unmanned vehicle are disclosed. The method comprises: establishing an object model through a reinforcement learning method, wherein the object model includes: a state of the unmanned vehicle, an environmental state described by a map picture, and an evaluation index of a path planning result; building a deep reinforcement learning neural network based on the object model established, to obtain a stable neural network model; inputting the map picture of the environment state and the state of the unmanned vehicle into the deep reinforcement learning neural network after trained, and generating a motion path of the unmanned vehicle. According to the present disclosure, the environment information in the scene is marked through the map picture, and the map features are extracted through the deep neural network, thereby simplifying the modeling process of the map scene.

Abstract: A multi-screen display system and a mouse switching control method are disclosed. The mouse switching control method is applied to a multi-screen display system comprising a main display screen and at least one extended display screen, and comprises: obtaining user images collected by cameras installed on the main display screen and the extended display screen respectively; inputting the user images into a neural network model, and predicting a screen that a user is currently paying attention to using the neural network model to obtain a prediction result; and controlling to switch a mouse to the screen that a user is currently paying attention to according to the prediction result. The system and mouse switching control method are based on self-learning of visual attention, predict the current screen operated by the user, automatically switch the mouse to the corresponding screen position, and improve the user experience.

Abstract: A plane detection method and device based on a laser sensor are disclosed. The method includes: acquiring data of the laser sensor after starting detection; inputting the data into a detection model trained in advance, wherein the detection model is obtained by training with data corresponding to a medium type selected in advance and is capable of recognizing the medium type selected; judging whether an object to which the data belongs is a plane, and if the object is a plane, determining the medium type of the plane; and setting corresponding optimization methods for different medium types, and optimizing the data according to the medium type. The laser sensor recognizes the medium type by the machine learning model, and optimizes the two-dimensional laser data according to the recognition results, and thus forms a more refined map and performs more accurate positioning based on the two-dimensional laser data.

Abstract: The present disclosure discloses a path planning method and device and a mobile device. The method comprises: collecting environmental information in a viewing angle by a sensor of a mobile device, processing the environmental information by using an SLAM algorithm, and constructing a grid map; dividing the grid map to obtain a plurality of pixel blocks, using an area constituted of pixel blocks not occupied by obstacles as a search area for path planning, and obtaining a processed grid map; determining reference points by using pixel points in the search area, and deploying topological points on the processed grid map according to the reference point determined and constructing a topological map; and calculating an optimal path from a starting point to a preset target point by using a predetermined algorithm according to the topological map constructed. The present disclosure improves path planning efficiency and saves storage resources.

Abstract: A method and device for extracting key frames in simultaneous localization and mapping and a smart device. The method includes acquiring an image frame from an image library storing a plurality of image frames of an unknown environment, and performing feature extraction on the image frame to obtain information of feature points, wherein the information includes a quantity of feature points; acquiring relative motion information of the image frame relative to the previous key frame, and calculating an adaptive threshold currently used by using the relative motion information; and selecting a key frame according to the information of feature points and the adaptive threshold indicating space information of image frames.

Abstract: An image data processing method is provided for a mobile terminal. The method includes receiving by a first electronic device first image data of an environment collected by a second electronic device; determining one or more motion parameters of the second electronic device based on the first image data; determining a latency between a moment the first image data being transmitted by the second electronic device and a moment the first image data being received by the first electronic device; compensating the first image data based on the one or more motion parameters of the second electronic device and the latency as determined, to generate second image data; and displaying the second image data through the first electronic device.