273results about "Image feed-back" patented technology

A robot in accordance with at least some embodiments of the present technology includes a neck region through which the robot is configured to transmit and receive light. The robot defines a robot height and includes a body, a head carried by the body, and a neck extending between the head and the body along the robot height. The robot further includes a structural neck frame at a periphery of the neck in a plane perpendicular to the robot height. The structural neck frame defines windows distributed around the periphery of the neck. The robot further includes a sensor disposed at least partially within the structural neck frame. The sensor includes a laser configured to transmit light via the windows and a detector configured to receive light via the windows. A field of view of the sensor extends at least 320 degrees around the neck.

A cabin virtual docking method based on actually measured point cloud data belongs to the technical field of spaceflight docking, and comprises the following steps: acquiring original point cloud data of a cabin structural member; inputting the initial point set into a feature extraction network; the key features comprise decoding features and coding features, the coding features are obtained through the feature processing block, and the decoding features are obtained through the feature recovery block; inputting the extracted key features into a feature segmentation network, segmenting the key features by the feature segmentation network according to categories, and performing parameter fitting on the segmented key features; optimizing matching indexes of the key characteristics after parameter fitting to obtain pose adjustment amounts of the upper cabin section and the lower cabin section so as to realize alignment of butt joint end surfaces and axis superposition of positioning pin holes; and virtual docking between the cabin sections is carried out by using the pose adjustment amount. According to the method, the cabin surface data are directly obtained through measurement, error accumulation is avoided, accurate calculation of the initial relative poses of the two cabins is achieved, and an efficient and reliable solution is provided for high-precision butt joint.

In one aspect, a grid map construction method includes: acquiring a point cloud of a target scene in the current field of view; according to the point cloud in the current field of view, constructing an occupation grid map and a height grid map corresponding to each other in the current field of view; and then updating a target grid map of the target scene according to a point cloud occupation probability value of each grid in the occupation grid map and a point cloud height value of each grid in the height grid map in the current field of view.

The invention discloses a road culvert inner wall damage image detection and traffic capacity evaluation method and system, and relates to the technical field of road engineering detection and computer vision, and the method comprises the steps: collecting multi-source data through an annular camera, a laser section scanner and a mileage encoder carried by a detection robot, and carrying out the panoramic expansion and time-space registration; using a multi-scale attention panoramic segmentation network to identify damages such as lining cracks, joint slab staggering, damage and chipping, steel bar exposure and water leakage; calculating the clearance size and the deposition thickness of the culvert by adopting an ellipse fitting method; according to the method, the culvert detection speed reaches 50 m / h, and the damage detection accuracy reaches 89% or above.

The invention relates to the technical field of artificial intelligence, and discloses a disordered workpiece grabbing method based on three-dimensional visual perception and intelligent decision, and the method comprises the steps: collecting the color image data and depth image data of a scene, and generating the three-dimensional point cloud data of the scene based on the internal reference of a camera; performing target detection and instance segmentation processing based on the color image data to generate target segmentation mask data, calculating an outlier degree score and a top degree score of a target in combination with depth data, and selecting an optimal capture target according to a comprehensive score; through an intelligent target selection strategy based on YOLOv11-SAMV3 fusion, in combination with an outlier degree score and a top degree score, a grabbing target is comprehensively evaluated, an isolated workpiece located at the top of a stack is preferentially selected, the shielding and collision risks are reduced, intelligent and automatic grabbing target selection is guaranteed, and the grabbing efficiency and the system adaptability are improved.

The invention relates to an impregnated paper production quality detection method and system based on deep learning, and the method comprises the following steps: S1, obtaining an industrial camera image flow, sensor and process parameters, and event and work order information, and constructing a synchronous data packet; s2, preprocessing the synchronous data packet to obtain preprocessed image data, process data and an ROI mask; s3, obtaining a detection frame set and a segmentation mask set based on multi-model AI detection and segmentation according to the preprocessed image data, the process data and the ROI mask; s4, segmenting the mask set according to the detection frame set, and performing track clustering along the paper feeding direction to form a defect event flow; s5, on the basis of the defect event stream, performing online quality prediction and root cause analysis, and obtaining future interval quality prediction and key factor sorting; and S6, according to future interval quality prediction and key factor sorting, adaptive process optimization is carried out, and an optimal control instruction is obtained. According to the invention, the detection accuracy and fineness are effectively improved, and intelligent identification and cross-frame tracking of defects are realized.

The invention provides an intelligent control method and system for independent grabbing of a multi-station cooperative soft package lithium battery, relates to the technical field of automatic control, and aims at achieving the independent grabbing of the soft package lithium battery by detecting abnormal information generated when a transfer mechanical arm grabs the soft package lithium battery at a preset intersection point and responding to the abnormal information. And selecting and activating a local exception handling scheme from a pre-stored local exception handling scheme library, controlling the transfer mechanical arm to execute a grabbing adjustment action or a safe retreat action according to the scheme, and controlling the grabbing mechanical arm to switch a preset motion path to an avoidance path or a temporary hovering path. And after the transfer mechanical arm completes abnormal processing, the grabbing mechanical arm and the transfer mechanical arm are coordinated to recover normal production operation, and the problems of grabbing failure, workpiece damage, mechanical arm collision and the like caused by infinitesimal displacement or rotation in the high-speed transfer process of the soft package lithium battery in the prior art are effectively solved.

The invention relates to the technical field of intelligent detection and automatic control, in particular to a coal moisture content intelligent sensing and dust suppression control method based on hyperspectrum, which comprises the following steps: acquiring an image and sorting a gray interval, removing a high-anti-interference area, segmenting numbered blocks, counting a gray direction reversal identification interference area, extracting peak positions of each wave band, and calculating difference value groups. And aggregating the continuous image blocks to extract gray boundaries, matching the boundaries and spraying serial numbers to adjust and control, and outputting the moisture content and a dust suppression image. According to the method, a region segmentation method driven by reflection density is introduced, gray level warping and interference contour elimination are combined, a block set is constructed to enhance the processing pertinence, through multi-band peak position difference value and symbol sequence continuity judgment, peak position change grouping is established to identify the spectrum trend, boundary aggregation is completed in the gray level mean value difference value direction, and finally, the spectrum trend is identified. And a control chart is generated by combining serial number position information, and linkage of an image structure, spectral distribution and control feedback is realized.

The invention discloses a defect detection method and system for a silicon wafer based on visual inspection, and relates to the technical field of visual inspection, the surface image of the silicon wafer is deeply excavated based on a multi-scale attention mechanism, and the primary defect content of the silicon wafer is subjected to multi-level fusion to output a plurality of primary defect areas; and each primary defect region and the corresponding associated region determine a final defect region in semantic dimension analysis, so that the accuracy of the final defect region is further improved. Performing multi-stage analysis on the plurality of silicon wafer defect characteristics along the gray gradient and the morphological parameters to determine the defect grade of each silicon wafer defect characteristic; according to the technical scheme, the plurality of defect combinations are determined according to the feature form of each silicon wafer defect feature and the corresponding defect level, the corresponding visual report is determined based on the identification of each defect combination, and the defect regulation and control instruction of the silicon wafer is determined along the multi-level iteration of the plurality of visual reports, so that the accuracy of the defect regulation and control instruction of the silicon wafer is improved.

The invention discloses a steel plate cutting method and device, a medium and equipment, and belongs to the field of cutting. According to the method, three-dimensional coordinates of all points on the surface of the head of a steel plate are generated by obtaining real-time motion parameters of the steel plate and the head distance and angle detected by a millimeter-wave radar and combining a preset coordinate conversion model. And then, converting the three-dimensional coordinates into a steel plate head shape image by using a morphological reconstruction model constructed based on preset three-dimensional point cloud data and morphological feature extraction rules. According to the steel plate cutting method and device, accurate positioning and efficient cutting are achieved, the production efficiency and the material utilization rate are improved, and the problem that the steel plate cannot be accurately cut in the prior art is effectively solved.

The invention provides a nuclear fuel assembly grabbing positioning and deviation correction method and system based on industrial vision. The method comprises the steps of obtaining multi-view image data of a target nuclear fuel assembly and structured light deformation image data under a corresponding view; extracting an outer contour boundary of the target nuclear fuel assembly in a three-dimensional space based on the multi-view image data and the structured light deformation image data to obtain initial outer contour boundary data; performing pose positioning based on the initial outer contour boundary data in combination with original design data of the target nuclear fuel assembly to obtain a current pose parameter of the target nuclear fuel assembly in a three-dimensional space; performing deviation analysis based on the current pose parameter in combination with a preset standard pose parameter to obtain current deviation data and a corresponding deviation analysis result; and if the deviation analysis result is correctable deviation, performing instruction generation based on the current deviation data to obtain a pose correction instruction. The problem that the correction effect is poor in an existing method is solved.

The invention relates to an automatic welding seam extraction and path planning system for a welding robot. The automatic welding seam extraction and path planning system comprises a three-dimensional model analysis module, an automatic welding seam extraction and feature calculation module, a welding key point and welding gun posture generation module, a path planning and reachability verification module and a data output and state management module. The welding seam automatic extraction and feature calculation module automatically identifies a welding seam based on a geometric topological relation and extracts geometric and topological features of the welding seam; the welding key point and welding gun posture generation module generates a welding path point, an arrival point, a departure point and a touch locating point according to the welding seam characteristics, and calculates a corresponding welding gun space posture; the path planning and reachability verification module performs multi-constraint verification on the executable performance of the welding path; and the data output and state management module is used for uniformly managing and outputting the welding path and the multi-state judgment result thereof. The system is high in welding path performability, high in adaptability to complex structures, high in system expandability, high in engineering efficiency and high in practicability.

The invention provides a port container loading and unloading positioning method, system and equipment and a medium, and belongs to the technical field of port automation and machine vision positioning. The method comprises the following steps: acquiring a container surface image and synchronous sling motion data; the jitter amplitude is evaluated, the image is combined to analyze the illumination condition, and the multispectral image is adaptively selected or fused to generate an anti-interference image; angular point coordinates are extracted, the angular point coordinates and motion data are input into a vision-inertia tight coupling optimization model together for state estimation, and the real-time pose is solved; driving a lifting appliance-load dynamic model by using the motion data, and predicting pose deviation generated by inertia and swinging in the future; and the feedback control quantity based on the real-time pose and the feed-forward compensation quantity based on the prediction deviation are fused, and an accurate control instruction for driving the lifting appliance is generated. The problems of image quality degradation and feature loss caused by strong light, weak light and sling shaking are solved, and high-precision and high-robustness container positioning and alignment in a complex dynamic environment are achieved.

The invention relates to the technical field of image processing, in particular to a bogie part image recognition and positioning method and system based on deep learning, and the method comprises the steps: constructing a bogie part and bolt image data set, and carrying out the preprocessing and marking; a recognition positioning model (introducing an ECA attention mechanism, optimizing a BiFPN feature fusion structure and designing a special bolt detection head) is constructed based on an improved YOLOv8 algorithm, after the model is trained and optimized, images are collected through a multi-view camera and input into the model to obtain two-dimensional coordinates, the two-dimensional coordinates are converted into three-dimensional space coordinates in combination with camera calibration and binocular vision, and the three-dimensional space coordinates are obtained. And after post-processing optimization, transmitting to a robot control system to guide full-automatic decomposition. According to the method, the recognition accuracy and positioning precision of the bogie parts and the bolts in the complex environment are improved, the bolt positioning error is within + / -0.5 mm, the recognition accuracy is larger than or equal to 98%, and support is provided for intelligent bogie maintenance.

The invention discloses a point cloud-based suction pose calculation method and device and a medium, and the method comprises the steps: collecting point cloud data corresponding to a to-be-sucked article in real time, and carrying out the preprocessing of the point cloud data, and obtaining node data and adjacent data corresponding to the node data; acquiring a graph neural network model; space geometric information is introduced into a message passing and attention calculation mechanism in a graph convolution operator to calculate and construct an SSFConv layer, a feature extraction layer in a graph neural network model is replaced by the SSFConv layer, and an absorption pose calculation model is obtained; and inputting the node data and the adjacent data into a suction pose calculation model for calculation to obtain a suction pose corresponding to the to-be-sucked article. According to the invention, the accuracy and robustness of the suction pose can be improved, so that the suction effect is improved.

The utility model provides a kind of analytical modeling method of model measuring point three-dimensional coordinates and Euler angle, the utility model is suspended under crossbeam by crane by scanning workpiece, solve the deformation problem of large thin-walled part different from its working posture due to the workpiece to be scanned by gravity, scanner is installed at the end of manipulator, realize all-around scanning to workpiece, according to the distance signal of scanner to control the scanning distance between scanner and workpiece to be scanned, solve the fatigue of operator long time handheld scanner, while can stably keep scanner and workpiece to be scanned directly optimal scanning distance, the utility model can satisfy the automatic scanning of different size sheet metal parts in whole vehicle reverse engineering, solve the deformation problem of large sheet metal part different from its working posture due to gravity when being placed on traditional scanning platform, while three-dimensional coordinates and Euler angle of measuring point in whole vehicle crash test and body-in-white modal test can be obtained quickly, improve scanning efficiency, shorten reverse engineering cycle etc.

A method of adaptive inspection includes receiving data characterizing one or more images of an inspection region of an industrial machine acquired by an inspection system operating based on a first set of operating parameters. The inspection region includes a site feature. The method also includes determining, by an analytical model, one or more characteristics of the inspection region from the received data characterizing the one or more images of the inspection region. The method further includes generating a control signal based on the one or more characteristics of the inspection region and / or a user input. The inspection system is configured to perform a new inspection of the inspection region based on the control signal.

The application provides a laser etching device and method for large-format and large-size samples, and aims to solve the technical problem that the laser processing mode of linkage of the fast axis and the slow axis of the existing galvanometer causes the non-uniform processing speed of different pattern segments due to the speed difference of the fast axis and the slow axis, and forms the splicing trace with inconsistent response speed when facing complex filled patterns. The laser etching method for large-format and large-size samples provided by the application is realized by the laser etching device for large-format and large-size samples composed of a laser scanning part, a sample movement part, a bracket and a controller, and the method of combining "trajectory decomposition + multi-layer filling" is adopted to realize the laser etching processing without splicing traces, thereby avoiding the splicing trace problem caused by the inconsistent speed response of the traditional linkage system of the fast axis and the slow axis, and ensuring the reliability and consistency of the large-area etching processing.

The application discloses a cotton bale detection method based on multi-modal data timestamp synchronization and GPU / CPU collaborative acceleration, collects multi-modal data, marks timestamps, and performs GPU accelerated distortion correction on cotton bale images; the multi-modal data is synchronized and integrated according to timestamps; a main program adopts Nodelet combined with C++ multi-threading technology to process the reception of multi-modal data and the publication of multi-modal sensor coordinate systems in parallel, the main program running optimizes the transmission efficiency of multi-modal data; a cotton bale detection model is constructed, the optimized multi-modal data are input for training, the cotton bale detection model after training is optimized using TensorRT, and cotton bale images are input for inference, cotton bale target recognition data are obtained and transmitted to a CPU for post-processing operation, and OpenMP is used to accelerate point cloud traversal in the main program to match the cotton bale target recognition data with laser radar point cloud data; the matched multi-modal data are subjected to point cloud feature and plane fitting calculation, and the rectangular frame, pose information and score of the cotton bale target are obtained.

The application relates to autonomous laser weed eradication, and provides a system for damaging or killing plants, the system comprising: a first camera, a second camera, a light source, a control system, and a computing system, the computing system is configured to: receive, at a first time, a first image of at least one plant in the field captured by the first camera; identifying a plant in the first image; predicting the position of the plant based on the first image; causing the second camera to capture a second image of an area of the field including the predicted position; predicting a target position of the plant in the second image at a second time after the first time, wherein movement of the second camera relative to the surface during the elapsed time is taken into account; causing the control system to direct an optical path of the light beam toward the predicted target position; the light source is caused to emit a light beam toward the predicted target location of the plant for a length of time sufficient to damage or kill the plant.

The invention discloses a smudginess detection method, device and equipment, and the method comprises the steps: carrying out the detection operation of a ground image obtained in real time, and obtaining one or more local regions; for each local area, respectively acquiring height information corresponding to each point in the local area based on parallax of stereoscopic vision; and respectively executing a screening operation in each local area, and determining one or more points of which the height information meets a preset range in the local area. According to the technical scheme, the ground points meeting the preset height are detected, and dirt detection can be efficiently and reliably carried out in real time in various complex scenes.

The invention provides a printed book and periodical automatic cutting management method and system based on image recognition, and belongs to the technical field of image recognition, and the method specifically comprises the steps: determining a quality inspection processing strategy of a cutting result of an image recognition model according to adjustment data and time sequence distribution data of an adjustment time period, and carrying out quality inspection processing on the cutting result by utilizing a quality inspection processing strategy, determining abnormal cutting data under different adjustment modes according to a quality inspection processing result when it is determined that management and control processing of maintenance of the printing mechanism needs to be carried out according to the quality inspection processing result and the correlation degree of the adjustment data between the adjustment time periods, and carrying out management and control on the maintenance of the printing mechanism according to the quality inspection processing result. And based on the abnormal cutting data and the change data of the adjustment mode of the cutting mechanism, the management and control processing method of the maintenance processing of the printing mechanism is determined, and the reliability degree of the cutting processing is improved.

The invention relates to the technical field of automobile part oiling, in particular to a self-adaptive control automobile part precise oiling method and system.The self-adaptive control automobile part precise oiling method comprises the steps that a part oiling abrasion model set is built, and a target oiling abrasion model is determined in the part oiling abrasion model set according to a target part model; a part shooting unit is used for carrying out part surface defect identification on a to-be-oiled part to obtain a surface defect area, surface defect analysis is carried out based on the surface defect area to obtain a part surface wear value, a part internal wear value is obtained, the part surface wear value and the part internal wear value are combined to obtain part defect data, and the part defect data is obtained. And obtaining an optimal oil film thickness group according to the part defect data and the target oiling wear model, and oiling the to-be-oiled part based on the optimal oil film thickness group and the oiling device. The precision of the automobile part lubricating effect can be improved, and the part abrasion risk caused by improper lubrication is reduced.

A system and method are provided for scanning a sensor using an area of a vision system in conjunction with an encoder or other motion knowledge to capture accurate measurements of an object greater than a field of view (FOV) of the sensor. It identifies the features / edges of the object, tracking from one image to another, providing a lightweight approach to process the overall extent of the object for dimensioning. Logic automatically determines whether the object is longer than the FOV, resulting in a series of image acquisition snapshots occurring while the moving / conveyed object remains within the FOV and until the object is not within the FOV. At this time, the acquisition is stopped, and the individual images are combined into segments in the overall image. These images may be processed to derive the overall size of the object from the input application details.

A system for detecting a falling object on a table includes a camera unit configured to obtain at least one image of a surface of the table, and a server configured to detect, from the at least one image, a first state in which an object is on the surface of a table or a second state in which no object is on the surface of the table, obtain information, when the first state is detected, on a first time during which the first state continues, detect that a falling object is on the surface of the table when the first time is more than a first reference time, and detect that no falling object is on the surface of the table when the first time is the first reference time or less.

To provide a system and method for provisioning computing resources using computer vision.SOLUTION: A method comprises: receiving an image comprising a plurality of visual elements; and recognizing the visual elements within the image as objects that represent network elements. The recognized objects have network element types such as a database, a load balancer, a sub-network, and a virtual machine instance and comprise one or more of these types of network elements. The method also comprises: recognizing an arrangement of the recognized objects that represents a network architecture; and provisioning computing resources corresponding to the objects by provisioning network elements that comprise one or more of the network element types. The provisioning generates a network architecture for the network elements based on the recognized arrangement.SELECTED DRAWING: Figure 17

A method for detecting an obstacle, applied to a self-moving robot, including: transforming obstacle information into depth information; converting the depth information into a point cloud map, and determining coordinate data of a reference point on the obstacle; determining a valid analysis range in a height direction in the point cloud map; and determining, based on the coordinate data of the reference point, whether an obstacle is present within the valid analysis range.

The invention relates to the technical field of false eyelash image recognition, in particular to a gluing line defect detection method and system based on multispectral image recognition, and the method comprises the steps that false eyelash bases are attached to a heating aluminum ring, the false eyelash bases are connected in series through a silk thread, a gluing mechanism is used for gluing along the silk thread to form a gluing line, and the gluing line is connected with the heating aluminum ring; the gluing line is deformed by heating the aluminum ring, and eyelash is promoted to fill gaps of the gluing line; the platform deck rotates the glued false eyelashes to a detection area, the multispectral imaging mechanism obtains multispectral images of the top face and the single side face of a gluing line, the images are preprocessed, and the preprocessing on the images comprises the steps of eliminating ambient light interference through a spectrum correction algorithm and highlighting a defect area through a gluing line characteristic wave band operation enhancement algorithm; and integrating spectral image features by using a multi-source image fusion algorithm, generating a fusion feature map according to the preprocessed image, and inputting the fusion feature map into a convolutional neural network model to realize the detection of the defects of the gluing line with the radian.

The application discloses a slotting machine rigidity dynamic adjustment system and method based on material image recognition, relates to the technical field of slotting machine rigidity dynamic adjustment, and comprises the following steps: acquiring a material surface and a subsurface image sequence of a machining area of a slotting machine; inputting the image sequence into a pre-trained deep visual analysis model to output a material characteristic parameter set in real time; converting the material characteristic parameter set into a target rigidity value based on a rigidity demand mapping model; obtaining an initial rigidity adjustment amount according to the target rigidity value and generating a rigidity adjustment instruction to be sent to a rigidity adjustment execution mechanism for dynamic adjustment within a preset control window period before a tool contacts a target machining area; and fusing feedback data and image analysis results to correct the initial rigidity adjustment amount during the cutting process. The application realizes dynamic, accurate and predictable active adjustment of the rigidity of the slotting machine, thereby effectively improving the machining adaptability, intelligence and stability of the slotting machine.