6 results about "3d tracking" patented technology

Systems and methods are described, and one system includes a three-dimensional (3D) geometric tracker connected to a multiband (MB) inverse synthetic aperture array radar (ISAR), and a classification / alarm logic. The MB ISAR includes spatially distributed radar transmitters (TXs) and receivers (RXs), a TX / RX allocation logic, and a tomographic (TM) image logic. The TX / RX allocation logic is configured to receive 3D tracking data from the 3D geometric tracker, indicating subject 3D position and 3D orientation and, in response, dynamically allocate TXs and RXs to maintain MB illumination of and maintain MB reception of multiple scatter angles from subjects. The TM image processor is configured to construct TM images from the scatter angles, using 3D tracking data, for input to the classification and alarm logic. Optionally, the TX / RX resource allocation logic is configured to receive situation feedback data, for feedback adjusting of allocation of TXs and RXs.

This invention discloses a method for real-time 3D tracking and localization of microscale targets in a bright-field microscopic environment, relating to the fields of image recognition and reconstruction technology. The method includes inputting continuous microscopic images into a target detection model, outputting the target's 2D position, regressing the defocus height and direction from a single frame of local images using a fine-grained ROI-based defocus regression model, directly using the model output to drive a 2D motion platform and a focusing device, fusing the 2D position information of each frame with the corresponding defocus axis information, and incorporating the position information of the previous frame as a priori using a temporal correlation mechanism to achieve continuous correlation of the microscale target in the time dimension, thereby stably reconstructing the target's 3D position and outputting a smooth real-time motion trajectory. This invention achieves real-time 3D localization and stable tracking of microscale targets under single-view bright-field microscopic conditions, ensuring both positioning accuracy and system real-time performance and accuracy.

In various examples, a technique for multiple object tracking is disclosed that includes generating, using one or more processing units, one or more first encoded image features based on a first image. The technique also includes generating a plurality of first object embeddings based on the first encoded image features, wherein at least one first object embedding of the plurality of first object embeddings corresponds to a different object depicted in the first image. The technique further includes determining, using a first track query of one or more track queries and one or more machine learning operations, a first association between a first track and a first object that corresponds to the at least one first object embedding, wherein the first track corresponds to the first track query. The technique further includes computing, based on the first association, an object trajectory associating the first track with the first object.

A controller (210) for tracking an interventional medical device (252) in three dimensions includes a memory (212) that stores instructions, and a processor (211) that executes the instructions. When executed by the processor (211), the instructions cause the controller (210) to execute a process. The process includes determining (S320 / S420), based on an elevation plane in an ultrasound X-plane mode, a first two-dimensional location of the interventional medical device (252) in the elevation plane. The process also includes determining (S320 / S422), based on an azimuthal plane in the ultrasound X-plane mode, a second two-dimensional location of the interventional medical device (252) in the azimuthal plane. The process moreover includes determining (S330 / S430), based on the first two-dimensional location and the second two-dimensional location, a three-dimensional location of the interventional medical device (252). Finally, the process includes modifying (S340 / S440) ultrasound beam patterns fired in the ultrasound X-plane mode based on the three-dimensional location of the interventional medical device (252).

The invention relates to the technical field of recognition and positioning, and particularly discloses a multi-target recognition and positioning method and system for a robot, and the method comprises the steps: firstly converting discrete three-dimensional point cloud data into a dense depth map aligned with a two-dimensional video frame pixel level in real time through a camera and radar equipment; based on this, the image is dynamically segmented into a plurality of depth regions, and a customized slice and scale transformation strategy is generated for each region. And further, far and near targets with great size differences in the original image are respectively converted into image slices with proper sizes for processing. And finally, unified detection and result inversion are performed on the scale-normalized slices, and three-dimensional tracking is performed in combination with the pose of the robot, so that robust recognition and accurate positioning of the multi-scale target are realized, visual scale dramatic change caused by the change of the distance of the target can be actively eliminated, and the positioning accuracy of the target is improved. Therefore, robust recognition and accurate positioning of the multi-scale target are realized.

This invention relates to a 3D tracking method for sea ice instances based on promptable VGGT (Virtual Geometry-Groograph) technology. The method includes: acquiring a sequence of polar sea ice images; performing 3D geometric encoding on the polar sea ice image sequence to obtain a 3D structure embedding vector for the scene; acquiring prompt text; performing language encoding on the prompt text to obtain a prompt semantic embedding vector; inputting the 3D structure embedding vector and the prompt semantic embedding vector into a cross-attention calculation module to constrain the 3D feature information under the prompt semantics; performing 3D geometric decoding on the 3D feature information under semantic constraints; and combining this with a temporal correlation module to achieve cross-time-sequence instance consistency modeling, generating a segmentation mask and instance ID; and performing instance-level fusion of the point cloud based on the segmentation mask to finally obtain the 3D tracking information of the sea ice. This invention can flexibly specify the tracking object and achieve accurate tracking of sea ice of different scales and morphologies.