1817 results about "Visual system" patented technology

A display system of an aircraft includes a display device defining a display area, a vision system for generating an image of an operating environment of the aircraft on the display area, data processors operatively coupled to the display device and the vision system, and non-transitory machine-readable memory operatively coupled to the one or more data processors and storing instructions executable by the one or more processors and configured to cause the one or more processors to, when a condition of the aircraft is met, cause an alteration of a display of flight information on the display area to declutter a cutout region in the image generated by the vision system, the cutout region devoid of imaging generated by the vision system and displaying an Approach Lighting System (ALS) of a runway of intended landing for the aircraft.

The invention relates to the technical field of visual automatic calibration, and discloses a joint manipulator automatic calibration method and device based on a visual system. The method comprises the following steps: carrying out multi-view image acquisition on a joint manipulator of the five-axis robot to obtain original image data; performing occlusion region extraction on the original image data to obtain a joint occlusion region set; on the basis of the joint occlusion region set, disordered image splicing and mark point feature extraction are carried out on the original image data, and joint feature image data are obtained; constructing an adaptive calibration equation based on the joint feature image data, and solving a target transformation relation between a global visual coordinate system and a local visual coordinate system; multi-joint collaborative calibration and joint chain constraint optimization of the five-axis robot are executed according to the target transformation relation, and joint error correction parameters are generated, high-precision joint error correction parameter calculation is achieved, and the positioning precision and the movement precision of a joint manipulator of the five-axis robot are greatly improved.

A dynamic vision system includes a variable focus liquid lens optical assembly. The dynamic vision system includes a variable lighting assembly. The dynamic vision system includes a control system configured to adjust one or more optical parameters and data collected from the variable focus liquid lens optical assembly in real time. The dynamic vision system includes a control system configured to adjust the variable lighting assembly. The dynamic vision system includes a processing system that dynamically learns on a training set of outcomes, parameters, and data collected from the variable focus liquid lens optical assembly to train a set of machine learning models to control the variable focus liquid lens optical assembly to optimize collection of data for processing by the set of machine learning models.

The invention provides a moving target capturing method and system based on AI multi-view vision, and relates to the technical field of data processing.The method comprises the steps that camera parameters in all vision systems are calibrated, and the camera parameters comprise a camera internal parameter matrix and a camera external parameter matrix; extracting 2D feature points of the moving target through each calibrated visual system; reconstructing each 2D feature point by combining a multi-view geometric reconstruction algorithm and a least square method according to the calibrated camera parameters to obtain 3D feature points; constraints are added to the 3D feature points, and the constraints comprise spatial constraints, direction vector constraints and time sequence constraints; reconstructing the 3D feature points added with the constraints through a multi-view geometric analysis algorithm to obtain three-dimensional motion information of the moving target; and outputting the three-dimensional motion information of the moving target to complete the capture of the moving target. And non-contact high-precision capture of the moving target can be realized.

The invention is suitable for the technical field of laser welding, and particularly relates to a laser welding control method which comprises the steps that an initial polarization image and an initial thermal imaging image of a to-be-welded workpiece are obtained; according to the initial polarization image and the initial thermal imaging graph, the welding path and initial laser parameters of the to-be-welded workpiece are determined; based on the welding path and the initial laser parameters, a laser welding device is controlled to weld the to-be-welded workpiece, and meanwhile a real-time polarization image and a real-time thermal image of the to-be-welded workpiece are obtained; according to the real-time polarization image, the real-time thermal image and the welding path, real-time state information of a molten pool of the to-be-welded workpiece is determined; and real-time laser parameters are determined according to the real-time state information of the molten pool, and the laser welding device is controlled to weld the workpiece to be welded according to the real-time laser parameters. According to the method, accurate welding process control is achieved, and the defects of a traditional visual system are overcome.

The invention relates to the technical field of automatic control of a dispensing process, and discloses an intelligent dispensing control method based on a self-adaptive track and dynamic fault-tolerant control. The method comprises the following steps: acquiring spatial pose data of a workpiece to be dispensed through a visual system, identifying a Mark point, and generating positioning reference information; the information is fused with height measurement data, and a dispensing track and dispensing height control parameters are obtained through calculation; driving an injection mechanism to dispense glue according to the control parameters, collecting a glue dispensing image, and comparing the image template to realize real-time judgment and alarm control of the offset and blockage state of the nozzle; after dispensing is completed, coverage rate analysis is conducted on the dispensing image, and a qualification result used for workpiece circulation judgment is generated; all control parameters, image data and judgment results are uploaded to a manufacturing execution system, the time interval between dispensing and the previous process is recorded, and process tracing and time card control are achieved. According to the invention, the dispensing precision and consistency are improved, and the automatic and fault-tolerant performance is good.

The invention provides a target cooperative control method and system based on deep learning, and the method comprises the steps: obtaining the real-time visual data of a target region through a multi-view image collection device, carrying out the recognition and classification of object features through a deep learning model, and generating feature distribution data containing a class identifier and boundary coordinates, a space calibration algorithm is adopted to map the two-dimensional features into three-dimensional coordinate system data, a space position and posture angle parameters are output in combination with posture calculation, a control instruction including path planning, tail end executor posture adjustment and self-adaptive clamping force is generated according to physical attribute parameters of an object, an operation device is driven to complete the precise grabbing action, and the grabbing precision is improved. Therefore, through dynamic multi-view perception, three-dimensional space modeling and physical attribute driven closed-loop control, the problems of feature deficiency, positioning errors, operation rigidity and the like of a traditional visual system in a complex scene are effectively solved, and the precision and adaptability of target operation in an unstructured environment are remarkably improved.

The invention is suitable for the technical field of electrical equipment corrosion detection, and provides an electrical equipment corrosion detection method based on double-spectrum collaborative analysis and space-time prediction, and the method comprises the following steps: carrying out fusion analysis on a visible light image and an infrared image of acquired electrical equipment, and outputting position information and morphological category of a corrosion region; a binocular vision system and an IMU sensor are combined to calculate three-dimensional space coordinates, main direction angles and environmental parameters of the corrosion area, and pose data are generated; inputting the pose data into a preset corrosion extension path prediction model for processing, and outputting physical characteristic parameters; fusing the dual-frequency RTK GPS positioning data, the IMU attitude data and the physical characteristic parameters to construct a space-time association data set, and inputting the space-time association data set into a preset space-time diagram neural network model to analyze a space-time evolution relationship of diagram structure modeling corrosion nodes; and outputting microscopic and macroscopic prediction results. The identification accuracy of corrosion characteristics and the corrosion detection efficiency can be improved, and the labor cost and the operation risk are reduced.

The invention discloses a full-automatic grinding method and system for steering knuckle flash based on 3D vision, and relates to the technical field of casting post-processing automation. A trimming-free direct polishing process is innovatively provided, a casting is scanned in real time through a high-precision 3D visual system, a flash space molded line is accurately captured, and one-key calibration of a camera, automatic generation and dynamic simulation of a flash path, multi-tool intelligent switching and self-adaptive adjustment of residual height are achieved in combination with self-developed off-line programming software. According to the system, a robot is adopted for grabbing workpieces to cooperate with a fixed type multi-cutter work station, a disordered feeding technology is integrated for 2D / 3D visual guiding grabbing, workpiece deformation and die difference are effectively dealt with, and flash residues are thoroughly eliminated. The method is characterized in that visual scanning, path planning and execution control are seamlessly fused, the dependence of traditional automatic grinding on the trimming process is broken through, the yield and efficiency are remarkably improved, and the industrial problem of full-process automation of casting flash is solved.

The invention relates to a stereoscopic warehouse checking method and system based on an unmanned aerial vehicle, and the method comprises the steps: receiving a cargo checking task, and obtaining the layout information of a stereoscopic warehouse; determining the position of a to-be-checked goods location in the goods checking task in the stereoscopic warehouse; according to the layout information of the stereoscopic warehouse and the position of the to-be-checked goods location in the stereoscopic warehouse, determining a safe flight area and a hovering point of the unmanned aerial vehicle based on a path planning algorithm, and generating a flight task; issuing a flight task to the unmanned aerial vehicle, and calculating pose information of the unmanned aerial vehicle during execution of the flight task in real time through an inertial measurement unit; based on the observation image of the unmanned aerial vehicle visual system, correcting the accumulated error of the pose information through a Kalman filtering algorithm; rGB color images and depth images collected by the unmanned aerial vehicle at the to-be-checked goods locations are obtained in real time; determining whether the box stack is in an abnormal condition according to the depth image; and for the box stack which is not in the abnormal working condition, the number of goods in the box stack is determined according to the RGB color image.

The invention belongs to the technical field of artificial intelligence and computer vision, particularly relates to a visual processing edge adaptive decision-making method for resource and task perception, and aims to solve the problem of scheduling mismatch caused by resource dynamic change and task demand diversity in visual task processing in an edge computing environment. The method comprises the following steps: collecting multi-dimensional resource state data of edge nodes in real time to form a resource state vector with high time resolution; analyzing the visual task request, and constructing a quantifiable task feature vector; and establishing a resource-task association mapping model based on a dynamic weight distribution mechanism. The method also supports cross-edge domain collaborative decision, and processes a pipeline dynamic reconstruction and security isolation mechanism. According to the technical scheme, the fluctuation of the resource utilization rate is reduced to 15% or below, the average task processing delay is reduced to 60%, the scheduling satisfaction degree is improved by 40% or above, and the self-adaptability and the service quality guarantee capability of the edge vision system are remarkably enhanced.

The invention discloses an interferometer error calibration method and device and a computer readable storage medium, and belongs to the technical field of measurement equipment.The method comprises the steps that measurement data collected when an interferometer moves on the x / y axis of a motion table is obtained, and x / y axis compensation amount is generated through calculation processing; meanwhile, a coordinate system transformation matrix is established and optimized by using data of a marking plate arranged on a motion table, and x-axis and y-axis non-orthogonal parameters are obtained. And then the compensation amount and the non-orthogonal parameters are input into a preset kinematic model, and the accurate position of the motion platform under the global coordinate system is calculated in real time. And when a preset updating condition is met, dynamically updating the error compensation value of the interferometer through the deviation between the actual position detected by the visual system and the calculated position. Reflector surface type error compensation and non-orthogonal parameter correction are fused, and high-precision error compensation calibration of the interferometer is achieved.

The invention relates to the technical field of constructional engineering, and discloses a bridge hanging basket installation accurate positioning automatic operation method, which comprises the following steps of: acquiring image data of the position, the speed, the posture and the construction environment of a bridge hanging basket in real time by installing a high-precision sensor and a visual system; analyzing the acquired image data by using a deep learning model, and identifying and estimating the spatial position and angle of the hanging basket; the system comprises a high-precision sensor and an image acquisition system which are used for acquiring the position, attitude and environmental data of the hanging basket in real time; the deep learning module comprises a convolutional neural network and a recurrent neural network and is used for recognizing the position of the hanging basket in real time and predicting future deviation; and the nonlinear optimization module is used for calculating optimal control input in real time according to the optimization objective function and correcting the optimal control input. By adopting the technical scheme of the high-precision sensor and the image acquisition system, real-time accurate monitoring of the position, the posture and the construction environment of the hanging basket is achieved.

This invention overcomes disadvantages of the prior art by providing a vision system and method of use, and graphical user interface (GUI), which employs a camera assembly having an on-board processor of low to modest processing power. At least one vision system tool analyzes image data, and generates results therefrom, based upon a deep learning process. A training process provides training image data to a processor (optionally) remote from the on-board processor to cause generation of the vision system tool therefrom, and provides a stored version of the vision system tool for runtime operation on the on-board processor. The GUI allows manipulation of thresholds applicable to the vision system tool and refinement of training of the vision system tool by the training process.

The invention, which relates to the technical field of engineering automation, discloses a BIM-based cable bridge routing autonomous optimization visualization system comprising a heterogeneous professional semantic modeling module, a path identification module, a path generation module, a path conflict evolution module, a path reconstruction module and an output module. The system identifies a cable thermosensitive area by constructing a cross-professional semantic map, constructs a multi-target path evaluation function and generates a high-quality candidate path by adopting a variable weight genetic algorithm; introducing a construction stage dynamic compression prediction mechanism to improve the path robustness; visual expression of the whole process of path design is realized through versioned path reconstruction and a three-dimensional animation evolution function; and finally outputting structured path data and an obstacle avoidance strategy which can be used for construction guidance. According to the invention, the intelligent level of cable bridge path planning, the engineering implementability and the system integration efficiency are significantly improved.

The invention discloses an outdoor distribution box inspection system and method based on path planning. The system comprises a master control terminal, an inspection unmanned aerial vehicle, an unmanned aerial vehicle landing base station and a distributed RTK positioning module. Wherein the main control terminal integrates an image processing unit and a dynamic path planning engine, and interacts with the patrol unmanned aerial vehicle in real time through a wireless communication protocol; the patrol unmanned aerial vehicle carries a dual-mode visual system and comprises a vertically downward path planning camera and an acquisition camera for direction adjustment of a holder, and the holder dynamically adjusts the orientation of the acquisition camera based on real-time position feedback of the RTK module to ensure that the optical axis of the holder is always aligned with the RTK module at the top of the nearest distribution box.

The invention provides an intelligent dancing garment generation method and system based on multi-modal action analysis. Dance movement biomechanical data are collected through an inertial sensor array and a multi-view visual system, features are extracted through a space-time diagram convolutional network, style semantics are analyzed in combination with a CLIP model, and a design drawing is generated through a diffusion model and fused into physical constraint optimization. And binding the 3D model with the virtual digital human to simulate a dynamic effect, and outputting a production instruction containing the fabric, the model and the process parameters. The system comprises a data acquisition module, an action analysis module and the like, and the whole-process intelligentization is realized. According to the method, traditional limitation is broken through, by fusing biomechanical data and artistic style semantics, the tear strength and style matching accuracy of the clothes are improved, the design period is shortened, automatic design and production of the dancing clothes are achieved, the dynamic adaptability and artistic expressive force of the clothes are improved, and an efficient scheme is provided for customization of the dancing clothes.

The invention provides a visual system, which comprises an element group and a lens cone group, when the diopter of the visual system is from + 2D to-5D, the moving distance delta L of the second element group along the optical axis, the combined focal length fz1 of the reflective polarizing element, the first quarter-wave plate and the first lens, and the combined focal length fz2 of the second lens, the second quarter-wave plate and the polarizing film meet the condition that delta L * (fz1 / fz2) is greater than or equal to 0.90 mm and less than or equal to 1.82 mm; and the maximum thickness Lb from the first side surface of the second lens barrel to the second side surface of the second lens barrel along the direction of the optical axis, the outer diameter Dbs of the first side surface of the second lens barrel and the inner diameter dbs of the first side surface of the second lens barrel meet the condition that Lb / (Dbs-dbs) is more than or equal to 0.56 and less than or equal to 2.36. The problem that smooth zooming and image quality improvement of a visual system in the prior art are difficult to consider at the same time is solved.

The invention discloses a compliance control method for a massage robot with a rigidity self-adaption function. The massage robot is composed of a six-degree-of-freedom mechanical arm, an end effector, a six-dimensional force sensor, a control host and other modules. The six-dimensional force sensor is installed at the tail end, the contact force of the robot and the surface of the human body can be collected in real time, tail end motion data obtained through the pose sensor is combined, the system executes a compliant control closed loop, and self-adaptive and safe massage operation is achieved. According to the method part, a force control closed loop can be constructed in real time, accurate, safe and smooth massage operation can be achieved through a force sensor installed at the tail end of the robot in combination with a smooth control algorithm based on displacement errors without depending on a visual system, meanwhile, massage area rigidity detection and self-adaptive control are achieved, and the user experience is effectively improved.

A computer-vision system or engine that (a) generates from a pixel stream a digital representation of a person and (b) determines attributes or characteristics of the person from that digital representation and (c) based on those attributes or characteristics, outputs data to a cloud-based analytics system that enables that analytics system to identify and also to authenticate the person. The attributes or characteristics of the person include their pose, and the system or engine analyses that pose to extract a facial image from a video stream that is the best facial image for use by the cloud-based analytics system to identify and authenticate the person. The computer-vision system or engine outputs the facial image to the cloud-based analytics system, but does not output the full-frame real-time video to the cloud-based analytics system.

The invention relates to the technical field of visual identification, and provides an AI visual system for shaft part stain identification and cleaning equipment, an image acquisition module acquires a surface image of a shaft part, generates an image of the shaft part, captures the surface image of the shaft part in real time, and provides a data basis for subsequent stain identification; the stain identification module processes the shaft part image, identifies the local contrast of each shaft sub-block in the shaft part image to obtain a stain candidate area, detects whether a stain edge exists in each shaft sub-block in the stain candidate area, combines the shaft sub-blocks to obtain a stain area, and triggers an equipment signal; the real-time monitoring and identification capability enables the production process to be more efficient, and problems can be found and processed in time. Corresponding cleaning operation can be automatically executed according to equipment signals of the visual system, the shaft parts are cleaned, manual intervention is reduced through the automatic process, and the production efficiency and the product quality are improved.

The invention provides a VR visual system, the VR visual system comprises an element group and a lens cone group, the element group sequentially comprises a first element group and a second element group from a first side to a second side along an optical axis of the VR visual system, the second element group can move along the direction of the optical axis, the combined focal length fz1 of the reflective polarizing element, the first quarter-wave plate and the first lens and the inner diameter dam of the second side surface of the first lens barrel meet the condition that fz1 / dam is more than or equal to 1.60 and less than or equal to 2.66; and the outer diameter Dbs of the first side surface of the second lens cone, the inner diameter dbs of the first side surface of the second lens cone and the distance delta L of the second element group moving along the optical axis when the diopter of the VR visual system is from + 2D to-5D meet the following condition: (Dbs-dbs) / delta L is greater than or equal to 2.63 and less than or equal to 7.45. The VR visual system solves the problem that a VR visual system in the prior art is large in diopter adjustment range and is difficult to consider the structural stability at the same time.

The invention discloses a road target detection method based on brain-like dual-channel feature fusion, and belongs to the technical field of image recognition. According to the method, a target recognition and spatial positioning dual-channel mechanism based on a human brain vision system is provided, a traffic scene video stream is converted into a pulse density map, the pulse density map is input into a bionic dual-channel network, feature extraction is performed from a target recognition channel and a spatial positioning channel, then adaptive dynamic feature fusion is performed through a dynamic fusion module, and a traffic scene video stream is obtained. By strictly mapping a human brain visual path, enhancing the theoretical reliability of the technical scheme and introducing pulse convolution, high real-time performance and calculation efficiency are ensured, the precision, speed and power consumption of traffic target detection are balanced, and the method is suitable for a vehicle-mounted traffic target detection task and meets the requirement of an automatic driving system for target detection. Experiments prove that the mAP of the method for small target detection is improved by 15.8%, and the energy consumption is reduced by 40%.

The invention discloses a robot vision external parameter calibration method based on blind deblurring and natural vibration measurement, and the method comprises the steps: collecting a continuous image sequence in a vibration environment through a high-frame-rate camera, extracting pixel-level motion information in combination with an advanced image deblurring deep network and an optical flow feature tracking technology, and carrying out the self-vibration measurement. And further restoring a three-dimensional displacement signal caused by natural vibration of the camera based on a pinhole imaging model, and obtaining ideal position information through frequency domain analysis to realize high-precision visual system external parameter calibration under a vibration condition.

The invention discloses a stacked workpiece disordered grabbing method and system based on a 3D visual system, and relates to the technical field of computer vision, and the method comprises the steps: obtaining an RGB image and a depth image of a stacked workpiece through a depth camera, and generating point cloud data; the improved Mask R-CNN model integrated with the MSDA module is used for performing instance segmentation on RGB and depth images, and workpiece pose estimation is realized by combining an ICP algorithm and template point cloud registration; establishing a coordinate relation between the camera and the mechanical arm through hand-eye calibration, and obtaining a 6D pose of the workpiece under a mechanical arm base coordinate system; and according to the point cloud centroid distance and the point number, the grabbing priority is calculated, the optimal target is selected for grabbing, and disordered grabbing of all the workpieces is completed circularly. According to the method, the improved Mask R-CNN model integrating the MSDA module and the ICP algorithm are combined with the 3D visual system, so that accurate instance segmentation and pose estimation of stacked workpieces are realized, and therefore, an unordered grabbing task is accurately completed.

The invention discloses a self-adaptive grabbing mechanical arm system for an automatic production line, and particularly relates to the technical field of self-adaptive grabbing, a target object is analyzed by using an artificial intelligence deep learning algorithm through a visual system and a tactile system, and the grabbing force and posture of a mechanical arm are automatically adjusted based on the characteristics of the target object. The method comprises the following steps: acquiring mechanical characteristic information and position information at the moment of contacting a target object, continuously monitoring and acquiring attitude characteristic information after contacting the target object, and constructing a visual risk assessment model based on a BP neural network by using the mechanical characteristic information and the attitude characteristic information; and taking the output of the visual risk assessment model as a target function of visual closed-loop feedback of the mechanical arm, and obtaining a simulation grabbing position corresponding relation according to the process that the mechanical arm grabs different positions of a target object in an automatic production line and based on the output of the visual risk assessment model. And the grabbing success rate and stability of the mechanical arm are improved.

The invention relates to the technical field of unmanned aerial vehicle vision, and discloses a vision system of an unmanned aerial vehicle and the unmanned aerial vehicle. The system comprises a visual data acquisition module, a dynamic feature extraction module, an environment modeling module and a decision control module. The visual data acquisition module captures a synchronous frame sequence containing infrared wave bands, visible light wave bands and depth information in a target area through a multispectral sensor array; the dynamic feature extraction module performs cross-modal fusion processing on the original visual data stream to generate a space-time correlation feature tensor containing a target contour geometric invariance descriptor and a motion trail differential topological structure; the environment modeling module constructs a three-dimensional semantic grid map according to the feature tensor, wherein each voxel unit codes the material reflectivity, the dynamic obstacle occurrence frequency and the illumination attenuation coefficient; the decision control module generates a flight path control instruction containing a pitch angle adjustment amount, a yaw angle compensation value and a speed change gradient based on the map, and assists the unmanned aerial vehicle to better cope with a complex environment.

The invention belongs to the technical field of robot positioning and three-dimensional mapping, and discloses a high-precision map reconstruction method and system based on monocular vision, a medium and equipment. Scene image data are acquired through a monocular image acquisition module, after feature extraction and matching are performed on each frame of image, attitude information acquired by an inertial measurement unit is fused, and pose calculation of a robot is completed through a sparse vision SLAM system; meanwhile, an image dense depth map is generated by a monocular depth estimation model based on an attention mechanism, and the image dense depth map is converted into a single-frame color dense point cloud in combination with image RGB color information. According to the invention, a loose coupling fusion strategy is adopted to carry out spatial registration on a single-frame colored dense point cloud and a robot pose at a corresponding moment, multi-frame data fusion is completed through point cloud splicing and optimization, and a globally consistent three-dimensional dense point cloud map is constructed to realize scene modeling. The method has the characteristics of high robustness and high reconstruction precision, and can be effectively applied to three-dimensional map construction in an outdoor complex environment.

The utility model discloses a bionic service type humanoid robot head structure which comprises a head structure, a sensory system, a mouth system, a visual system, a nose system and an ear system, the mouth system, the visual system, the nose system and the ear system are connected with the sensory system, and the mouth system, the visual system, the nose system and the ear system are installed on the head structure. The mouth system comprises a jaw structure, a tooth structure, a lip structure and a tongue structure, the sensing center, the sensor network, the signal processing unit and the execution unit form a sensory system, the sensory system can timely and effectively feed back external environment changes, and the capability of interacting with the outside is greatly improved; meanwhile, according to the taste sensor, the humanoid robot can test and evaluate the food under the condition that no human testee exists, and the taste sensing system can simulate the function of a human taste sensing system, so that the taste of the food is automatically evaluated.