30 results about "Global vision" patented technology

The invention discloses a warehouse navigation intelligent vehicle scheduling method based on global vision. The warehouse navigation intelligent vehicle scheduling method comprises the steps of: 1) shooting a global image by using a top camera on a ceiling of a warehouse; 2) processing the global image to track intelligent vehicles in the warehouse, and identifying a number of each AGV by using convolutional neural network identification; 3) dividing the warehouse into a plurality of regions, and sending instructions to control the AGVs by using a global controller, so that the AGVs operate according to routes; 4) performing regional path planning by using an improved A* algorithm to obtain an optimal passing region set of the AGVs from a current region to a target region; 5) utilizing atime window algorithm to plan routes in the region so that the AGV can safely exit the current region; 6) and utilizing a segmented PID control algorithm to enable the AGVs to complete scheduled transportation tasks along the planned routes. The warehouse navigation intelligent vehicle scheduling method fully allocates AGV passage tasks to all the regions, and improves the operation efficiency ofan AGV storage system.

The invention relates to a SDN network DDoS and DLDoS distributed space-time detection system comprising detection nodes arranged in various SDN switches. Each detection node comprises a data acquisition module, a space-time anomaly detection module, and an output module. The data acquisition module is used for acquiring network flow flowing across the SDN switch. The space-time anomaly detection module is used for detecting the network flow acquired by the data acquisition module in a spatial domain, determining whether suspicious flow exists, and confirming whether DDoS or DLDoS exists in a time domain on the basis of a detection result of the spatial domain. The output module is used for storing the detection result according to a set format. In view of the characteristics of the DDoS and the DLDoS, the detection system detects and discriminates the DDoS and the DLDoS in the network flow spatial domain and time domain, and achieves global vision in abnormity detection by using a distributed virtual ANN overlay network.

The invention discloses a multi-warehouse AGV tracking method based on global vision, and the method comprises the following steps: 1), photographing a global image of a warehouse, and transmitting the global image to a control center; 2) processing the global image, enabling the tracking system to use a target detection algorithm to identify the plurality of AGVs in the first frame, and determining the ID and pose of each AGV according to the AprilTag code at the top of the AGV; 3) dividing the warehouse into a plurality of areas, and performing path planning of each AGV by using a hierarchical planning algorithm after the scheduling system acquires AGV pose information sent by the control center; 4) combining the path information of each AGV with a multi-AGV tracking algorithm, predicting the positions of the AGVs, selecting an area where the AGVs are located by using a boundary frame, and determining the information of each AGV; and 5) enabling the control center to send the speed instruction converted from the path information to the AGV, and controlling the AGV to complete a cargo sorting task.

The invention relates to a strategy regulatory module of a global vision soccer robot decision-making subsystem and a method thereof, wherein the module comprises a plurality of resource databases, such as a role library, a formation library, a basis graphics library, a robot action library, and so on. The module and the method of the invention can be used for re-drawing a competition area, assigning roles, modifying a role parameter and adding new role functions, and so on via graphical programming or code programming; especially the decision-making subsystem can be adjusted only on a strategy processing panel without entering into a complex program, which greatly strengthens the operability of the whole system. In addition to this, a strategy parameter can be directly modified in the game and taken effect in time so as to be convenient for an operator to timely modify our tactics according to different strategies of an opponent and a site situation.

The invention discloses a flexible assembly welding robot workstation that comprises a global vision unit, a flexible welding robot group, a gantry truss detection unit, a master control unit and a conveying platform; the flexible welding robot group is arranged on the two sides of the conveying platform, welding workpieces are arranged on the conveying platform, the global vision unit is arrangedon the periphery of the flexible welding robot group, the gantry truss detection unit is arranged above the conveying platform, and the flexible welding robot group is controlled by the master control unit to conduct welding according to obtained image information. The high integration degree, high flexibility and high efficiency of the welding robot are achieved, the problems that manual assembly spot welding position deviation is large, and workpiece consistency is poor are solved, and the workpiece clamping and transferring frequency is effectively reduced. Multi-position assembly spot welding of workpieces is achieved, seamless connection of welding procedures is achieved, and the assembly precision, the welding consistency and the welding quality are improved.

The invention relates to a high-precision global vision measurement method for a three-dimensional thin-wall structure weld joint. The method comprises the following steps: (1) preparing a workpiece and a measurement system; (2) calibrating the measurement system and optimizing parameters; (3) measuring a space curve weld joint; (4) carrying out data splicing and welding seam extraction; (5) splicing a skin and blanking; and (6) calibrating a robot welding track. The method is innovatively applied to the field of measurement of the three-dimensional contour weld joint; on-site automatic measurement of the three-dimensional contour of the actual space curve weld joint is achieved, the weld joint position and the contour characteristics of the weld joint are recognized and extracted, on thisbasis, weld joint track calibration in the robot off-line programming process is achieved, and accurate discharging of the skin in the weld joint assembling process can be achieved.

The autonomous motion planning method of large-scale space manipulator based on multi-channel vision fusion. In multi-channel vision fusion, the priority of the global vision system is lower than that of the local vision system, and the priority of the local vision system is lower than that of the wrist vision system. The global vision system can obtain the pose information of the target object, but neither the local vision system nor the wrist vision system can obtain the pose information of the target object, so the target pose information obtained by the global vision system is used for motion planning; For the movement of the arm body, if the local vision system can obtain the pose information of the target object, before the manipulator body reaches the middle point of tracking, the target pose information obtained by the local vision system is used for motion planning; when the manipulator body moves to the tracking At the middle point, at this time, the wrist vision system has been able to obtain the pose information of the target object, and the motion planning is performed using the target pose information obtained by the wrist vision system. The invention realizes the tasks of autonomous movement, autonomous tracking and autonomous capture of target objects in a large space range.

The invention provides a global vision-based chess playing robot and a control method thereof, and relates to the technical field of computer vision. According to the global vision-based chess playingrobot and the control method thereof, a sucker is connected with an air pump and an actuator body correspondingly, and a tail end camera is connected with the actuator body and a computer server correspondingly; an anthropomorphic mechanical arm is connected with the computer server and the actuator body correspondingly, and the mechanical arm is located at the blank position in the middle of a chess piece taking disc; a view camera is connected with the computer server and is arranged on a camera support, and the computer server comprises a visual system, a game system, a chess piece placingsystem and a real-time display system; and the visual system is connected with the game system and the real-time display system correspondingly, the game system is connected with the chess piece placing system and the real-time display system correspondingly, and the output end of the chess piece placing system is connected with the mechanical arm. The global vision-based chess playing robot andthe control method thereof overcome the interference of moving objects, enhance the accuracy of chess piece positioning under uneven illumination, and create a scientific, convenient and fast chess playing environment for a user.

The invention discloses an aerial maintenance robot locating method and device based on near-earth global vision. By means of a double-vision system, the global position coordinates of a robot body and fault points are acquired through a first camera shooting module, and the robot body is controlled to carry out primary locating according to the obtained global position coordinates; and the localposition coordinates of the fault points are acquired through a second camera shooting module arranged at the tail end of a mechanical arm of the robot body, and the robot body is controlled to carryout secondary precise locating. According to the aerial maintenance robot locating method and device based on near-earth global vision, the first camera shooting module serving as a global camera is arranged at the near-earth end, and the technical problems that after the global camera ascends to a high altitude along with an insulation body, the camera bears the action of vibration of the mechanical arm during work and high-altitude wind power, camera disturbance is caused, then the precision of primary locating of the global camera is reduced, and the precision of secondary locating of the robot body is influenced is avoided.

The invention discloses a global vision positioning method and system for a small quadruped robot, and the method comprises the following steps: calibrating a dual-camera system, and calculating the motion model parameters of the dual-camera system; obtaining a calibrated target area image shot by the fisheye camera; processing the image of the target area to obtain the position of a to-be-positioned target in a fisheye camera coordinate system, calculating the spatial position of the to-be-positioned target in a holder coordinate system through a double-camera system kinematics model, and carrying out two-degree-of-freedom kinematics inverse solution to obtain a two-degree-of-freedom joint rotation angle; obtaining a to-be-positioned target image, detecting a two-dimensional code in the to-be-positioned target image to obtain a two-dimensional coordinate of a two-dimensional code center point in the to-be-positioned image, and performing transformation to obtain a two-dimensional coordinate of the two-dimensional code center point in a ground coordinate system; and correcting the two-dimensional coordinates of the two-dimensional code center point in the ground coordinate system to obtain the position and orientation of the to-be-positioned target in the ground coordinate system.

The invention relates to a soccer robot dual meet system based on global vision and remote collaborative control. The system comprises a global vision subsystem, an intelligent decision-making subsystem, a USB wireless communication subsystem, and a robot car subsystem. When the system is running, the global vision subsystem converts image information acquired by an image acquisition device into pose information of robots and position information of a ball and transfers the information to the intelligent decision-making subsystem, the intelligent decision-making subsystem judges field attack and defense state through an expert system according to the acquired information, assigns a role to each car automatically and obtains a command value of the speed of the left and right wheels through conversion, and the USB wireless communication subsystem sends the wheel speed command value to the corresponding robots in the field through a transmitting module to control the motion of the robots. Meanwhile, a participant can also wirelessly and remotely control the motion of any car on the side thereof through an offered remote control handle (only one car can be controlled at the same time), and the priority of a remote control command is higher than that of a command issued by the expert system.

Embodiments of the present application provide a visual attribute recognition method and device and a medium. The method includes acquiring basic visual information of a target image by using a firstpart of a preset neural network model; using the second part of the preset neural network model to obtain the global visual information and the local visual information; according to the importance ofeach feature in the local vision information, processing the local vision information to obtain the first local adjustment information; according to the importance of each feature in the global vision information, processing the global vision information to obtain the first global adjustment information; adjusting the global vision information according to the first local adjustment information to obtain the adjusted global vision information; adjusting the local vision information according to the first global adjustment information to obtain the adjusted local vision information, and performing the visual attribute recognition on the target image according to the adjusted global vision information or/and the local vision information. The accuracy of visual attribute recognition can be improved.

The invention discloses a cascade expansion method for a working space and a working visual angle of a stereoscopic vision system. One stereoscopic vision system in the multiple stereoscopic vision systems serves as a global vision system to be fixed, and the other stereoscopic vision systems can be randomly moved in position and adjusted in view angle according to requirements in the working process, so that the working space and the working view angle of the stereoscopic vision systems are effectively expanded, and the flexibility of the stereoscopic vision systems is improved. A to-be-measured target cannot be seen under a global stereoscopic vision system; indirect positioning of a to-be-measured target under a global coordinate system can be achieved through other stereoscopic visionsystems, the purposes of overcoming limited visual angle and limited working space of a single stereoscopic vision system are achieved, the limitation that poses of multiple stereoscopic vision systems in a traditional scheme cannot be adjusted in the working process is overcome, and application of the stereoscopic vision systems is wider. The method is applied to operating room instrument navigation and has the advantages of low cost, flexible working space, high positioning precision and the like.

The invention discloses an underwater robot image secret character embedding and decoding method. The method comprises the following steps: combining an original image I and to-be-coded information M1into four channels as input, and outputting a coded image O and a residual image R through a U-shaped network, so as to obtain a first loss; inputting the coded image O, enhancing the robustness andgeneralization ability of the model, and calculating a cross entropy to obtain a second loss; weighting the first loss function and the second loss function, and analyzing and obtaining a function model with strong coding capability and low perception loss; training a BiSeNet semantic segmentation network, generating a large number of coded images O, predicatign alternative image areas where the coded images possibly appear, and then putting the alternative image areas into a decoder to be detected and judged. According to the method, the deep learning steganography is applied to encoding, decoding and retrieval of the secret information of the bionic robot, global vision individual ID deep learning steganography mode identification, and encrypted and hidden watermarking of a consumer-level underwater robot.

The invention discloses a sequence data pixel visualization adaptive visual mapping adjustment method and system. The adjusting method comprises the steps of S100, conducting pixel visualization on multiple pieces of sequence data input by a user on a screen according to a preset mapping method and global vision mapping, wherein each sequence comprises multiple data points, and each data point corresponds to one pixel; S200, according to the size of the screen, setting a display area of the plurality of pieces of sequence data on the screen; and S300, after it is determined that the user selects a part of the display area on the screen, calculating a mapping range of the part of the display area, and adjusting global visual mapping including the part of the display area according to the mapping range. According to the invention, the visual mapping in the pixel visualization is automatically adjusted according to the mapping range of the partial display area selected by the user, so that the pixel visualization can better display the local features of the high dynamic range data.