14 results about "3d sensor" patented technology

This invention belongs to the field of physical testing, specifically relating to an automated detection system for surface defects on a 3D sensor protective film. The system includes an acoustic-thermal excitation device, a thermal imaging sensing unit, an optical coherence tomography (OCT) scanning device, and a data processing and analysis terminal. The acoustic-thermal excitation device emits a focused ultrasonic beam onto the protective film surface to generate an acoustic-thermal effect; the thermal imaging sensing unit captures the spatiotemporal attenuation characteristics of the surface temperature field and generates a two-dimensional thermal diffusion feature map; the OCT scanning device adaptively acquires a three-dimensional structural cross-sectional image of the protective film based on the hotspot location; the data processing and analysis terminal performs spatial alignment and pixel-level fusion of multimodal data, and uses a deep neural network to identify defects. Through the above technical solution, this invention achieves non-contact, non-destructive, high-precision detection, balancing detection efficiency and accuracy. It can accurately identify deep, minute defects under complex coatings and predict the impact of defects on optical performance.

The present invention provides a screw detection system capable of appropriately detecting the position and orientation of a male screw, a screw fastening system equipped therewith, and a screw detection method. [Solution] The screw detection system comprises a 3D sensor and a processing circuit. The processing circuit generates a brightness distribution image of male screws with different brightness levels depending on the distance from distance data, generates a second derivative image by taking the second derivative of the brightness distribution image, generates a third derivative image by taking the third derivative of the brightness distribution image, calculates the logical AND of the second derivative image and the third derivative image to generate an edge extraction image showing the brightness distribution at the top of the screw threads of the male screw, converts the brightness values ​​in the edge extraction image into 3D point cloud data for distance, performs a cylinder fitting process to fit a cylinder model with a preset outer diameter to the 3D point cloud data, and determines the position and orientation of the central axis of the cylinder model after the execution of the cylinder fitting process as the position and orientation of the central axis of the male screw.

Systems and methods detect scan avoidance behaviors at a self-service point-of-sale (POS) terminal. The system includes a camera positioned to capture video images of an operational area of the POS terminal and a computer that determines when movement of an item by a customer operating the POS terminal indicates scan avoidance. The computer receives video images from the camera, preprocesses frames of the video images to suppress background and to track the movement of the item from a pickup area of the POS terminal, through a scan area of a scanner of the POS terminal, and to a bagging area of the POS terminal. The system may also include a 3D sensor for detecting depth data that is used to improve hand-item contact.

This invention provides a crane system and a control method for the crane system that can reduce the load on data processing. [Solution] The crane system 1 comprises a moving mechanism for moving the hoisting machine 10, a position detection unit 20 for detecting the horizontal and vertical positions of the load P, a 3D sensor 30 for measuring the three-dimensional coordinates of obstacles that hinder the transport of the load P, a data processing device 100 for processing data related to the obstacle based on the coordinate data from the 3D sensor 30, and a control unit for controlling the operation of the moving mechanism and the hoisting machine 10 based on the data processing results in the data processing device 100. The data processing device 100 includes an inclusion rectangle processing unit 105 that creates inclusion rectangle data R corresponding to the rectangle data including the obstacle region H in the 3D data including the obstacle region H indicating the obstacle.

This invention discloses a 3D security sensor early warning and protection method based on deep learning, comprising the following steps: S1, acquiring a 3D sensor point cloud frame sequence; S2, constructing a spatial reachability function based on security protection rules, and labeling the reachable point set frame by frame in the point cloud; S3, performing KPConv kernel convolution under reachability constraints to extract 3D features restricted by security boundaries; S4, aggregating the 3D features by spatial units to generate a frame-level structure embedding vector sequence; S5, inputting the structure embedding vector into a DAGMM network, calculating the anomaly energy value and filtering anomaly candidates; S6, performing anomaly reversibility testing on the anomaly candidates, and obtaining the minimum perturbation cost through constrained reverse optimization; S7, based on the joint determination of the anomaly energy value and the minimum perturbation cost, outputting the early warning level and the set of early warning events. This invention achieves accurate identification and graded early warning of 3D security anomalies, reduces false alarms, and improves early warning stability.

The present invention relates to the field of optical measuring devices, and more particularly to displacement sensors, 3D sensors, or roughness sensors for measuring the position and / or shape or thickness of an object to be measured. Measuring light is projected onto the object to be measured after being filtered by a first Fabry-Perot filter, and at each point on the measuring surface, the filtered measuring light has a locally unique wavelength or combination of wavelengths in at least one direction within the measuring surface. Measuring light reflected from the surface of the object to be measured is filtered by a second Fabry-Perot filter to remove measuring light that has not been reflected from the intersection of the object to be measured and the measuring surface.

This invention discloses a method and system for pig body posture perception based on point cloud completion, belonging to the field of livestock and poultry intelligent perception and 3D point cloud processing technology. This invention acquires 3D point clouds of the pig's body surface from a single viewpoint using a 3D sensor. After denoising, background cropping, downsampling, and coordinate normalization, a standardized missing point cloud is obtained. The missing point cloud is input into a pre-trained Transformer encoder-decoder point cloud completion model. Through a feature interaction learning module combining local geometric and global structural features, and combined with an inverted residual multilayer perceptron module, high-precision completion of the missing regions is achieved. Based on the completed point cloud, 3D reconstruction of the pig's body posture is completed, and multi-dimensional analysis including body size measurement, health classification, and posture recognition is performed. The results are visualized and synchronized to a smart livestock management system. This invention adopts a single-view, low-cost deployment, adapts to the non-rigid deformation of pigs, has high completion accuracy, and strong anti-occlusion ability, making it suitable for pig health monitoring in large-scale smart livestock farming.

A system for detecting a position of a container door including: a 3D sensor arranged to detect and generate 3D spatial data map; said 3D spatial data map including an array of data points, said data points defined by their respective x,y,z coordinate, wherein x,y coordinate is a plane vertical to the 3D sensor axis and z coordinate a distance function from the 3D sensor perpendicular to the xy plane; a processor arranged to receive positional data from the sensor, said processor including an algorithm arranged to analyse the position data to determine the door closure status.

A method and a system for determining a path for processing a workpiece include causing a manipulator of an industrial robot to present a plurality of different postures; causing a 3D sensor mounted on the manipulator to capture point cloud data of the workpiece when the manipulator presents the respective different postures, one of the workpiece and the 3D sensor being mounted on the manipulator, the other of the workpiece and the 3D sensor being placed on a worktable; stitching the acquired point cloud data for the plurality of posture; and generating, based at least in part on the stitched point cloud data, the path for processing a workpiece.

A confocal three-dimensional sensor (22) for measuring height of a point on an object (10) is provided. The sensor (22) includes a light source (2) and a light source modulator (16) configured to temporally modulate the light source intensity. A source pinhole aperture (4) is positioned to be illuminated by the light source (2) and a focus-tunable lens (8) is configured to focus illumination passing through the source pinhole aperture (4) onto the object (10). A detector pinhole aperture (12) is configured to receive reflected light from the object (10), wherein the focus-tunable lens (8) is configured to image the reflected light from the object (10) onto the detector pinhole aperture (12). A detector (14) and integrator (15) are configured to output a measurement indicative of total light transmitted through the detector pinhole aperture (12). A processor (20) is operably coupled to the detector (14), integrator (15), and light source modulator (16). The processor (20) is configured to synchronously cause the light source modulator (16) to modulate light source intensity while causing the focus tunable lens (8) to sweep axial focal position, the processor (20) being further configured to calculate a height of the point on the object (10) based on the output from the detector (14) and integrator (15).

Systems and methods detect scan avoidance behaviors at a self-service point-of-sale (POS) terminal. The system includes a camera positioned to capture video images of an operational area of the POS terminal and a computer that determines when movement of an item by a customer operating the POS terminal indicates scan avoidance. The computer receives video images from the camera, preprocesses frames of the video images to suppress background and to track the movement of the item from a pickup area of the POS terminal, through a scan area of a scanner of the POS terminal, and to a bagging area of the POS terminal. The system may also include a 3D sensor for detecting depth data that is used to improve hand-item contact.

Method for detecting a leak in a fluid guiding element 16, 18 of a heat exchanging device 12, comprising the following steps:a) Acquiring a cloud of points 44 of the heat exchanging device 12 in a three-dimensional virtual space 48 using a 3D sensor 34, each of said points representing a surface point 46a on the outer surface of the heat exchanging device 12, b) Searching within the cloud of points 44 for structures 50 corresponding to tubes having a predefined outer diameter,c) Searching the obtained structures 50 for an area of interest 52 where the diameter or the direction of the tube changes,d) Once an area of interest 52 is found, determining an approach path 58 for approaching a sniffing probe 22 of a gas leak detector 24 to the area of interest 52 within the cloud of points 44, ande) Physically approaching the sniffing probe 22 to a testing area 20 of the heat exchanging device 12 automatically along the approach path 58, where the testing area 20 corresponds to the area of interest 52 within the cloud of points 44.

This invention discloses a method and system for identifying the illumination invariance of wild and farmed large yellow croaker based on cross-modal 3D shape priors. First, two-dimensional RGB images of the large yellow croaker are acquired, preprocessed, and then input into a network. The network includes a shared encoder, an auxiliary depth reconstruction branch, and a main classification branch. The auxiliary depth reconstruction branch uses an attention gate mechanism to reconstruct pixel-level depth maps from the RGB images, extracting illumination-invariant geometric features. The main classification branch uses a shape-guided feature fusion mechanism to fuse the geometric features with the texture features extracted by the shared encoder, achieving accurate classification between wild and farmed large yellow croaker. This invention also employs a semi-supervised balanced batch sampling training strategy based on privileged information, effectively utilizing a small amount of 3D data to guide the model in learning geometric priors. This invention can significantly improve the robustness and accuracy of the model under complex lighting conditions without increasing the cost of expensive 3D sensors.