30results about How to "Improve the quality of acquisition" patented technology

The invention provides a visual 3D taking and placing method and system based on a cooperative robot. The method comprises the steps that internal and external parameters of a camera of a binocular structured light three-dimensional scanner are calibrated; the hands and eyes of the cooperative robot are calibrated, and a calibration result matrix is obtained; a three-dimensional digital model of to-be-taken-and-placed target objects is collected; the calibrated binocular structured light three-dimensional scanner is adopted to obtain point cloud data of the to-be-taken-and-placed target objects which are stacked in a scattered mode, and the point cloud is segmented to obtain scene point clouds of the multiple to-be-taken-and-placed target objects; the to-be-taken-and-placed target object with the highest grabbing success rate is selected as a grabbing target according to the scene point clouds of the multiple to-be-taken-and-placed target objects; the three-dimensional digital model ofthe grabbing target and scene point pair features are registered, pre-defined taking and placing pose points are registered into a scene, and a registered pose estimation result is obtained and serves as a grabbing pose of the grabbing target; and a preliminary grabbing path track of the cooperative robot is planned. The target object can be accurately recognized, and the grabbing positioning precision is high.

The invention discloses a method and a device for performing content distribution based on edge wireless hotspots. The method is characterized by comprising the following steps that: S1, a content distribution network (CDN) region is partitioned into a plurality of sub-regions; S2, the wireless hotpots in the sub-regions collect video watching records of users in the sub-regions and periodically report the collected video watching records to a scheduling server, and the scheduling server predicts popular video content of the sub-regions in a future set time interval according to the video watching records; S3, a CDN server duplicates the popular video content to the wireless hotspots in corresponding sub-regions to store under the condition that a current bandwidth is smaller than a set bandwidth; and S4, the users check whether the video content is stored in the wireless hotspots or not when requesting the video content, if so, directly acquire the video content from the wireless hotspots, and otherwise request the video content from the CDN server. Through adoption of the method and the device, the user experience can be improved, and meanwhile the service overhead of a CDN system is reduced.

The invention belongs to medical instruments and wearable electronic products, and relates to a flexible semi-dry electrode which is of a multi-layer structure and used for collecting electroencephalogram. The electrode comprises a flexible electrode body layer, a foam layer and a liquid storage layer. A flexible probe in the flexible electrode body layer is of an array sensor structure; a plurality of through holes for the flexible probe to run through are arranged on the conductive foam cotton in the foam layer, and the plurality of through holes are arranged in accordance with the arrangement of the flexible probe; electrolyte is stored in the liquid storage layer and the liquid storage layer is provided with a release assembly for releasing the electrolyte to the conductive foam cottonin the foam layer. The flexible electrode has the advantages of convenient use, comfort, reasonable structure, low impedance, long-term stability and reliability, flexibility and comfort and safety of wearing.

The invention relates to a Chinese-Vietnamese parallel sentence pair extraction method based on cross-language bilingual pre-training and Bi-LSTM and belongs to the technical field of natural languages. The method comprises the following steps: firstly, collecting Chinese-Vietnamese comparable corpora, and extracting Chinese-Vietnamese parallel sentence pairs from the comparable corpora; adding aChinese-Vietnamese bilingual dictionary and a large number of Chinese-Vietnamese monolanguages in pre-training, performing word alignment by mapping the Chinese-Vietnamese bilingual dictionaries to apublic semantic space, then generating a new dictionary iteratively in a self-learning mode by the Chinese-Vietnamese seed dictionary so that semantic similarity between Chinese-Vietnamese sentences is represented to the maximum extent; then inputting Chinese and Vietnamese sentences obtained after pre-training into a twin neural network composed of Bi-LSTM and CNN, and extracting global featuresand local features of the sentences; and finally, judging whether the input sentence pair is a Chinese-Vietnamese bilingual parallel sentence pair or not by using a full connection layer. Good effectis achieved in an experiment of extracting parallel sentence pairs from comparable corpora.

A 3D face image generating method, device and electronic device based on occlusion image are provided in the invention. The method includes identifying target human face image in picture, judging whetthat rotation angle of the target human face is greater than a preset threshold value or not;; identifying feature point information in the target face image by a pre-trained regression model when the rotation angle is greater than a preset threshold value, the first feature point information including first visible feature point information and first invisible feature point information; the first feature point information is inputted to a pre-trained convolution neural network model to obtain a three-dimensional face image corresponding to the target face image. By using the method, the device and the electronic device, not only the problem of difficult generation of the three-dimensional face image due to the recognition and positioning of the feature points caused by the invisibility of the feature points due to the self-occlusion is solved, but also the effect of the generation of the three-dimensional face image is further improved, and the matching degree is improved.

The invention discloses an intelligent supplementary scanning method based on a two-axis turntable and a computer readable storage medium. The method comprises the following steps: S1, calibrating parameters of a binocular structured light three-dimensional scanner; S2, performing hand-eye calibration based on the two-axis turntable to obtain a calibration result matrix of a spatial relative position relationship between the scanner and a two-axis turntable base; S3, acquiring point cloud data of the three-dimensional digital model of the target object by using the calibrated scanner to obtain a point cloud model; S4, searching for a point cloud model missing area in a computer UI window, aligning a window view to the missing area, obtaining a point cloud model target pose displayed in the window in real time, and controlling a turntable to synchronously move according to the calibration result matrix and the target pose; S5, collecting a single point cloud under a target pose corresponding to the point cloud of the missing area, reconstructing the single point cloud, and registering the single point cloud with the point cloud data collected in S3 to complement the point cloud data of the missing area; and S6, repeating the steps S4 to S5 until all the point cloud data of the missing region of the point cloud model are completed, and acquiring a complete three-dimensional digital model of the target object for deviation detection.

The invention relates to the technology of cutter monitoring, in particular to a high-speed milling cutter abrasion vision in-place monitoring system and method. The system comprises a mechanical hand, a clamp, an industrial microscope lens and an industrial personal computer. The mechanical hand is fixedly installed on a ground foundation. The clamp is installed at the tail end of the mechanical hand. The industrial microscope lens is clamped on the clamp and electrically connected with the industrial personal computer, and obtained workpiece surface image data is transmitted to the industrial personal computer. The monitoring method comprises the steps that the equipment is prepared; the mechanical hand demonstrates; a standard block database is set up; and in-place detection is carried out. According to the system and the method, the industrial microscope lens obtains an image of the machined surface of a workpiece, as the milled surface is a plane, the problems of scattering and the like do not occur easily, the image obtaining quality is high, the surface quality data can be obtained accurately, and thus the abrasion condition of a cutter can be monitored accurately. The mechanical hand selects to carry out measurement during an interval of cutter stopping, and in-place detection is achieved in the cutter machining process.

The invention relates to an optical fingerprint recognition module, comprising: a casing, an image acquisition component, a light guiding diffusion layer, a light collecting reflective layer, a pressing plate, at least one light source and a light channel structure. The light channel structure is formed below the pressing plate and above the image acquisition component, passes through the light collecting reflective layer and a portion of the light guiding diffusion layer, and is surrounded by the light collecting reflective layer and a portion of the light guiding diffusion layer. The light emitted by the at least one light source can form conduction and diffusion in the light guiding diffusion layer, and can form condensation and reflection in the light collecting reflective layer, and the light emitted by the at least one light source can be reciprocated between the light guiding diffusion layer and the light collecting reflective layer, then irradiated onto the pressing plate via the light channel structure, and then reflected onto the image acquisition component to be acquired. The optical fingerprint recognition module allows the use of light to be more effectively increased.

The invention discloses a data processing method and device for questions and answers, equipment and a storage medium. The method comprises the steps of obtaining question and answer pair data; generating an initial question-answer pair set according to the question-answer pair data; splitting the initial question and answer pair set into a question set and an answer set; clustering the problem set, and extracting to obtain a problem cluster set; and generating a question cluster set with answers according to a preset processed question-answer pair set and the question cluster set. The technical problem that data used for questioning and answering is not well utilized is solved. According to the invention, the cost of manual knowledge construction is reduced. At the same time, the extracted hot question and answer pairs reduce the labor cost of compilation of knowledge by enterprises according to document terms or memory, and the generated knowledge points ensure the richness.

The embodiment of the invention discloses a picture processing method, a device, storage medium and a terminal. The method comprises the following steps: adopting a watermark removing instruction, wherein the watermark removing instruction indicates a picture needing to remove the watermark; sending a picture search request to the server according to the watermark removing instruction, wherein the picture search request carries the picture content of the picture, and the picture content is the content except the watermark content in the picture; acquiring a candidate picture of the picture content from the server; selecting the target picture not containing the watermark from the candidate picture, and saving the target picture. By using the scheme disclosed by the embodiment of the invention, the picture search can be performed through the server so as to acquire the watermark-free picture with the same content, namely, the original picture; the original picture can be acquired without importing the watermark picture to the dedicated watermark removing application and performing the complicated algorithm to remove the watermark; the original acquisition efficiency of the watermark picture can be improved, and the acquisition quality of the original picture is improved.

The invention discloses a method for measuring a volume of a rock-fill dam test pit through handheld three-dimensional laser scanning. The method comprises the following steps: firstly, leveling a dam charge level, placing a lantern ring, and excavating the test pit; then, according to the diameter and depth of the test pit and the stacking and filling material, determining the dot spacing, the laser color and the scanning mode; using a handheld color three-dimensional scanner to extend into the test pit, scanning according to the curvature of the pit wall from small to large, processing obtained point cloud: denoising, simplifying, packaging and filling, and then selecting a circle of annular area on the top of the test pit model for optimal plane fitting; finally, taking the fitting plane as an xoy plane, and enabling the z-axis direction to face downwards; and projecting the whole test pit model to the xoy plane, enabling the projection area to be an integral interval S, calculating the volumes of the small prisms of the test pit model, and adding to obtain the volume of the whole test pit. The method is flexible and simple to operate, and can quickly scan various types of test pits, obtain high-quality point cloud data, efficiently and quickly process the data and accurately calculate the volume of the test pit.

The embodiment of the invention provides a red Fuji apple shape data enhancement device and method, and the method comprises the steps: carrying out the photographing of an apple through a shape dataobtaining device under the control of a controller, and forming an original data set; performing classification calibration on the pictures in the original data set through a data classification device to obtain a calibration data set; dividing a region of interest for the pictures in the calibration data set through a data cutting device, cutting to eliminate noise, and reserving the region of interest to form a sample data set; and inputting the sample data set into a deep convolutional generative adversarial network through a picture generation device, performing training processing, generating a new apple picture, and forming a result data set to obtain enhanced data of the appearance of the red Fuji apple.

Examples of the present invention provide a method, device and electronic device for light information, which include: obtain the target picture, decompose the target picture into the first preset number of sub -pictures, where the sub -picture is from the sub -picture fromThe pixel composition of the second preset quantity; determine the light intensity of the sub -picture, and calculate the image torque to determine the light weighted center of the sub -picture;The geometric center determines the light angle of the sub -picture; calculates the light intensity and light angle of the sub -pictures, and obtains the light information corresponding to the target picture.There is no need to obtain the environmental information, location and time of pictures or videos, nor do you need to collect and label the face image or other specific hardware resources.In areas without light, this method can also be used to obtain light intensity.Compared with existing technology, the cost of the present invention is low, and it is more convenient and accurate.