45results about How to "Improve image recognition rate" patented technology

Extensive social communication is induced. Connection is made with a network terminal capable of connecting to the Internet, and an image and voice signal reflecting the subjective visual field of the user and the like which can be obtained from the headset system that can be worn by the user on the head is uploaded via the network terminal to a knowledge-information-processing server system, and specifying and selecting of an attention-given target by the voice of the user himself / herself are enabled on the server system with collaborative operation with the voice recognition system with regard to a specific object and the like to which the user gives attention and which is included in the image, and with regard to the series of image recognition processes and image recognition result made by the user, image recognition result and recognition processes thereof are notified as voice information to an earphone incorporated into the headset system of the user by way of the user's network terminal via the Internet by the server system with collaborative operation with a voice-synthesizing system, so that user's message or tweet can be extensively shared by users.

The present invention discloses an interactive presentation system which allows a presenter to perform a presentation while having various interactions directly with presentation material images in real time through a gesture or / and voice. The interactive presentation system comprises: an active infrared camera; a command recognition system connected to the active infrared camera; and an image synthesis system connected to the active infrared camera and the command recognition system. The presentation system may further comprises a stereo camera set for properly synthesizing a presenter in a 3D image and a 3D motion system. By this configuration, it is possible to embody an interactive presentation system in which a command through a presenter's gesture or voice is processed in real time and the image of the presenter is synthesized in a presentation material screen in real time, and accordingly the audiovisual effect is maximized.

The present invention discloses an interactive presentation system which allows a presenter to perform a presentation while having various interactions directly with presentation material images in real time through a gesture or / and voice. The interactive presentation system comprises: an active infrared camera; a command recognition system connected to the active infrared camera; and an image synthesis system connected to the active infrared camera and the command recognition system. The presentation system may further comprises a stereo camera set for properly synthesizing a presenter in a 3D image and a 3D motion system. By this configuration, it is possible to embody an interactive presentation system in which a command through a presenter's gesture or voice is processed in real time and the image of the presenter is synthesized in a presentation material screen in real time, and accordingly the audiovisual effect is maximized.

The invention discloses a space positioning device, system and method. The space positioning device comprises a communication unit for establishing communication with a positioning object; an image obtaining unit for obtaining a first stereo image of a background image under the condition that a mark point infrared light source of the positioning object is in an off state, and a second stereo image comprising a mark point under the condition that the mark point infrared light source is in an on state; an image processing unit for carrying out image processing on the first stereo image and the second stereo image to obtain a mark point light spot graph; and a positioning unit for obtaining a space coordinate of the mark point according to the mark point light spot graph. The space positioning device, system and method realize distinguishing between the object to be positioned and the background image, improve image recognition rate, reduce algorithm complexity, simplify algorithm operation environment and save cost.

The invention discloses an automatic inspection and real-time image acquisition transmission method of a railway line unmanned aerial plane. Through the method, automatic track plan inspection can be realized, visual navigation, image identification and obstacle avoidance are supported, parameters can be properly changed as no transmission delay influence is exerted, so the image identification rate is improved, a satellite communication network mode is fused with a traditional communication mode, high speed exchange for high-capacity image data can be realized, and the method further has an advantage of relatively high security.

The present invention relates to a food material input method and a food material input device for a refrigerator. The food material input method comprises: obtaining a food material image inside a storage room; identifying N kinds of predicted food materials and the confidence thereof according to the food material image, wherein N is a positive integer; and determining the type of the food material according to the N kinds of predicted food materials and the confidence thereof. In addition, the present invention also provides a food material input device by using the method. By virtue of thetechnical scheme of the present invention, in the food identification process, the accuracy of identification can be improved, the waste of resources can be reduced, and the user experience can be improved.

A vehicle safety control apparatus using cameras includes a first camera configured to photograph light and shade of a current object during running, a second camera configured to photograph a color of the current object during the running, an image processing unit configured to perform image processing on first current object image data captured by the first camera and second current object image data captured by the second camera, a recognition unit configured to recognize the first and second current object image data on which the image processing unit has performed the image processing, a storage unit configured to cause the recognized data to match preset reference object-specific data and separately store the data matching the preset reference object-specific data; and a control unit configured to receive the recognized data and deliver a storage command to the storage unit.

The invention relates to a feature weighted face identification algorithm, belongs to the field of human face technology. The feature weighted face identification algorithm comprises the steps that a face image is decomposed into high and low frequency components via wavelet transformation, principal component analysis (PCA) is carried out on the different components to extract feature images, weighing is carried out by an AHP algorithm according to the significance of the components, and a support vector machine (SVM) is used for classification and identification. The image identification rate is highest, and all main information of images is used for identification.

A vehicle safety control apparatus using cameras includes a first camera configured to photograph a current object at at least one of a narrow angle and a straight angle during running, a second camera configured to photograph the current object at a wide angle during the running, an image processing unit configured to perform image processing on first current object image data captured by the first camera and second current object image data captured by the second camera, a recognition unit configured to recognize the first and second current object image data on which the image processing unit has performed the image processing, a storage unit configured to cause the recognized data to match preset reference object-specific data and separately store the data matching the preset reference object-specific data; and a control unit configured to receive the recognized data and deliver a storage command to the storage unit.

The invention discloses a multi-purpose aircraft external store automatic mounting system and method, belonging to the technical field of automatic mounting. The system includes an image recognition module, a control solution module, a mechanical lifting arm execution unit, an integrated display control module and an external store transport vehicle. The method includes: at first, two frame of adjacent images collect by that camera in the image recognition module are processed differentially, and are transmitted to the position detection module for coordinate conversion, and the spatial relative position of the external store and the corresponding mount point under the wing are solved, and are transmitted to the control solution module as error control signals; and then, the relative position of the external store and the corresponding mounting point under the wing are converted to the coordinate system; the control solution module calculates the error to complete the control parametercalculation and generates the control signal to be transmitted to the mechanical lifting arm execution unit, and the mechanical lifting arm lifts the mounting platform away from the vehicle body to reach the target position and the target posture for operation. The invention effectively avoids the influence of environmental interference, improves the reliability and safety of the system, and greatly improves the intelligence and automation efficiency of the weapon equipment.

The embodiment of the present invention provides a method and device for protecting a camera against fog interference, and an electric appliance, and belongs to the field of electric appliances. The method includes: determining the lens temperature of the lens of the camera; determining the dew point temperature at the lens; and instructing a heating module to heat the lens according to the dew point temperature at the lens and the lens temperature. The device includes: a heating module for heating the lens of the camera; a lens temperature determination module for determining the lens temperature of the lens; a dew point temperature determination module for determining the dew point temperature at the lens; and a control module for instructing the heating module to heat the lens accordingto the dew point temperature at the lens and the lens temperature. Therefore, no fog is generated on the surface of the camera, or the fog is quickly removed after generated, thereby protecting the camera against fog interference and reducing the influence of the fog on the imaging effect of the camera.

The invention provides a method for obtaining the position of mine underground equipment based on character image intelligent identification, and the method comprises the steps: installing a pluralityof character boards beside an underground rail at intervals, marking a plurality of characters on each character board, and recording the rail position corresponding to each character number in a database of a production scheduling center; installing image acquisition equipment on the unmanned aerial vehicle under the mine, and acquiring all character images marked on the corresponding characterboards during driving; sending the acquired image to a U-Net network for detection; detecting an image with eight characters in the image; segmenting the image into eight non-overlapping sub-images, wherein each sub-image contains one character; after convolution operation and down-sampling operation, performing classification and identification to obtain corresponding character values and credible values, and transmitting the character value corresponding to the position with the maximum credible value to a production scheduling center in real time through a wireless network, so that the unmanned aerial vehicle can be accurately positioned underground, and industrial and automatic mining and transportation production requirements are met.

The invention discloses a high-power zoom unmanned aerial vehicle aerial high-definition multi-dimension real-time investigation transmitting method. The method comprises the following steps that firstly, a monitoring computer starts a monitoring program, and a satellite navigation unit starts a GPS navigation program; secondly, a high-definition high-power zoom action camera collects video images according to the track of the monitoring program, and a vehicle-side image processing module processes the images; thirdly, a video image wireless transmitting module and a video image receiving module cooperate to complete wireless sending and receiving of image signals; fourthly, a central station image processing module processes the received image signals, and the received image signals are displayed on a terminal. According to the method, visual navigation, image recognition and obstacle avoidance are supported, parameters can be properly changed under the situation that the influence of transmission delay is avoided, and thus the image recognition rate is increased; a satellite communication network mode and a traditional communication mode are fused together, high-capacity image data can be exchanged at a high speed, and high safety is achieved.

The invention provides a transformation-based X-ray image BGA (ball grid array) solder ball contour extraction method. The method includes steps: S1, acquiring a solder joint original image, subjecting the original image to circle detection to obtain a solder joint image, and subjecting solder balls of each solder joint to primary recognition to obtain ROI (region of interest) of each solder joint; S2, transforming the original image to obtain a target image, and subjecting the target image to edge detection; S3, mapping a detection result to an original coordinate system to obtain a mask image, and subjecting the original image to mask processing to extract images of single solder joints; subjecting the mask image to edge extraction to obtain a solder joint contour image.

The invention belongs to the technical field of image recognition, and particularly relates to an image recognition method based on a convolutional neural network. According to the method, different segmentation methods are adopted to perform image segmentation on the same image; and for different segmentation methods, different feature extraction algorithms are adopted to extract image features,the extracted image features are inputted to the same convolutional neural network model for identification, and the final identification result of the image is calculated in combination with the identification results of different weights.

The invention discloses a tree species image recognition method based on a natural background. The method comprises the steps of firstly capturing tree image data through employing a web crawler technology, storing the tree image data according to different types, and forming a data set; preprocessing the image, and segmenting a complex background in the original image from a target tree entity byadopting a Grab-Cut image segmentation method; learning the segmented image by using a convolutional neural network; adopting k-fold cross validation for model selection, wherein firstly, known dataare randomly segmented into K subsets which are not intersected with one another and are the same in size; training a model by using the data of K-1 subsets, and testing the model by using the remaining subsets; repeating the process for K times, finally selecting the model with the minimum average test error in K times of evaluation, and identifying the tree species image. According to the method, an image segmentation method is adopted, complex backgrounds which are not used for tree species identification are segmented, and the tree species identification rate is greatly improved.

The invention discloses a 3D biological printer spray head capable of detecting cell activity and a printing method, which is suitable for different printing models, has controllability, can meet the needs of high-precision biological printing, and can monitor cell activity in the printing process to make a judgment on whether the print is successful or not. The invention is realized by the following technical schemes: the 3D biological printer spray head capable of detecting cell activity includes a microscopic imaging system, a printing system and a slave computer control core which are arranged in a fixed housing, and the slave computer control core controls the microscopic imaging system to take pictures, store pictures and transmit pictures to a master computer at the same time; and the printing method using the 3D biological printer spray head includes two printing modes: a normal printing mode and a microscopic imaging mode.

An apparatus and methods for recognizing an image with an increased image recognition rate are provided. The apparatus includes a plurality of illuminators for illuminating a surface. The apparatus also includes an image sensor for outputting an image frame obtained by converting an optical image of the surface into an electrical signal. The surface is illuminated one of the plurality of illuminators. The apparatus further includes a switch for selecting any of the plurality of illuminators for illuminating the surface. The apparatus additionally includes a controller for receiving the image frame from the image sensor, determining whether the image frame has a region of reflected light, and controlling the switch to select another of the plurality of illuminators to illuminate the surface when the image frame has the region of reflected light.

The invention relates to a color image recognition method which is characterized by comprising the following steps: S1, acquiring a picture of a processing object; s2, performing channel selection processing; s3, executing subsequent processing for the selected specific channel; s4, extracting characteristic values of the characters according to a set recognition algorithm after processing is completed, comparing the characteristic values with a target value of the model, and calculating and judging a closest result; and S5, outputting identification result information. The method has the advantages that after the method is used, the original color information of the image can be utilized to the maximum extent, the contrast ratio and the discrimination degree are increased, information loss caused by gray processing is avoided, the recognition effect of the feature value is improved under the condition that other feature value recognition steps are not changed, and finally the image recognition rate is increased.

The invention discloses a man-machine voice interaction system and method for a shopping bin, and relates to the technical field of intelligent interaction. The method comprises the steps of samplingdata; acquiring multi-dimensional information by using user voice and a user identifier, and determining an identification model; allowing a user to collect a front-side image through photographing equipment; detecting the acquired image through face detection equipment, and carrying out human eye positioning; intercepting a face area in the image, and performing graying processing; storing the processed data; extracting SIFT features in the image to serve as training sample data to be stored in a data storage center; dynamically updating identification resources according to the division; anddetermining different user information of going in and out of the shopping bin by using a user voice recognition text. According to the man-machine voice interaction method for the shopping bin, thefunctions are diversified, the working efficiency is high, the maintenance cost is low, a new module can be added in later maintenance and updating to meet the requirements of a later system, and thefunctions of the man-machine voice interaction method for the shopping bin are more complete.

The invention discloses an automatic unmanned aerial vehicle aerial patrolling and real-time image collecting and transmitting monitoring device for a railway line. The monitoring device can automatically plan a patrolling track and support visual navigation, image recognition and obstacle avoidance, and parameters can be properly changed without transmission delay influences, so that the image recognition rate is increased; a satellite communication network mode and a traditional communication mode are fused together, high-capacity image data can be exchanged at a high speed, and high safety is achieved.

The invention discloses an imaging method and system for long-distance anti-atmospheric turbulence. Under the solar light field, a detector is used to perform N times of continuous imaging on a target object. The interval between two adjacent imaging is greater than the exposure time of the detector. The correlation between each point and itself, using the results of N consecutive imaging to calculate the intensity variance of each point, can obtain image quality with high contrast and signal-to-noise ratio. Compared with traditional direct imaging, the imaging scheme proposed by the present invention eliminates the influence of atmospheric turbulence, improves the practicability of associated imaging technology, can be directly used for long-distance anti-atmospheric turbulence imaging, and increases the image recognition rate.

The invention discloses an aerial high-definition multidimensional real-time detection and transmission monitoring device of a high-magnification zooming unmanned plane. The monitoring device comprises a monitoring computer, a plane-end image processing module, a satellite navigation unit, a high-definition high-magnification zooming moving camera and a video image wireless emission module that are installed on an unmanned plane. The plane-end image processing module includes a data receiving unit, a parameter changing unit, a decoding unit and an image identifying unit. The data receiving unit receives image encoding streams including image coded data and parameters. The parameter changing unit can change the parameters received by the data receiving unit. The decoding unit decodes the image coded data according to the parameters changed by the parameter changing unit, and generates image decoded data. The image identifying unit carries out image identification on the image decoded data that has been decoded. The device supports visual navigation, image identification and barrier avoidance, can properly change parameters without being affected by transmission delays, and thus the image identification rate is increased. A satellite communication network mode and a conventional communication mode are combined together, so the problem of high speed exchanges of high capacity image data is solved, and high security is exhibited.

The invention discloses an automatic building drawing information extraction method and system based on an image recognition technology, and the method comprises the steps: carrying out manual marking of a matching anchor point and a keyword character region of each building drawing template in advance, and obtaining a type matching template of a matching type template library; identifying the type matching template with the highest matching degree, and finding a plurality of component sub-matching templates matching the type template; extracting contour information of a to-be-recognized building drawing, performing component matching when each piece of contour information matches the sub-matching template of the component, and identifying component information of the to-be-recognized building drawing; or comparing each component sub-matching template in the component sub-matching templates with the contour of the architectural drawing to be identified so as to find the component information of the to-be-recognized architectural drawing; and intercepting the component information of the to-be-recognized architectural drawing, identifying the non-components of the to-be-recognized architectural drawing by using an SSD algorithm, and forming digital information of the to-be-recognized architectural drawing.

The invention discloses a high-power zoom unmanned aerial vehicle aerial high-definition multi-dimension real-time investigation transmitting system. The system comprises an image monitoring device installed in an unmanned aerial vehicle, and a video transmitting device installed in a ground central station. The image monitoring device comprises a monitoring computer, a vehicle-side image processing module, a satellite navigation unit, a high-definition high-power zoom action camera and a video image wireless transmitting module which are installed on the unmanned aerial vehicle. The vehicle-side image processing module comprises a data receiving unit, a parameter changing unit, a decoding unit and an image recognition unit. The data receiving unit receives an image encoding stream including image encoding data and parameters, the parameter changing unit can change the parameters received by the data receiving unit, the decoding unit decodes the received image encoding data and generates image decoding data according to the parameters changed by the parameter changing unit, and the image recognition unit carries out image recognition on the decoded image decoding data. The video transmitting device comprises a video image receiving module, a multi-channel distributing module, a central station image processing module and a display terminal. The image receiving module receives image signals transmitted by the image transmitting module.

The invention discloses an image processing method, which comprises: acquiring a first image, and determining a first target image and a second target image included in the first image; respectively identifying the first target image and the second target image to obtain a first identification result and a second identification result; and determining a target recognition result according to the first recognition result and the second recognition result. The invention further discloses a control method, a recognition method, a recognition device and a storage medium. The invention provides an image processing method for voice and image fusion, so that errors or deviations are adjusted, and the recognition accuracy is improved.

A vehicle safety control apparatus using cameras includes a first camera configured to photograph light and shade of a current object during running, a second camera configured to photograph a color of the current object during the running, an image processing unit configured to perform image processing on first current object image data captured by the first camera and second current object image data captured by the second camera, a recognition unit configured to recognize the first and second current object image data on which the image processing unit has performed the image processing, a storage unit configured to cause the recognized data to match preset reference object-specific data and separately store the data matching the preset reference object-specific data; and a control unit configured to receive the recognized data and deliver a storage command to the storage unit.