In recent years, satellites capable of capturing videos have been developed and launched to provide high definition satellite videos that enable applications far beyond the capabilities of remotely sensed imagery. Moving object detection and moving object tracking are among the most essential and challenging tasks, but existing studies have mainly focused on vehicles. To accurately detect and then track more complex moving objects, specifically airplanes, we need to address the challenges posed by the new data. First, slow-moving airplanes may cause foreground aperture problem during detection. Second, various disturbances, especially parallax motion, may cause false detection. Third, airplanes may perform complex motions, which requires a rotation-invariant and scale-invariant tracking algorithm. To tackle these difficulties, we first develop an Improved Gaussian-based Background Subtractor (IPGBBS) algorithm for moving airplane detection. This algorithm adopts a novel strategy for background and foreground adaptation, which can effectively deal with the foreground aperture problem. Then, the detected moving airplanes are tracked by a Primary Scale Invariant Feature Transform (P-SIFT) keypoint matching algorithm. The P-SIFT keypoint of an airplane exhibits high distinctiveness and repeatability. More importantly, it provides a highly rotation-invariant and scale-invariant feature vector that can be used in the matching process to determine the new locations of the airplane in the frame sequence. The method was tested on a satellite video with eight moving airplanes. Compared with state-of-the-art algorithms, our IPGBBS algorithm achieved the best detection accuracy with the highest F1 score of 0.94 and also demonstrated its superiority on parallax motion suppression. The P-SIFT keypoint matching algorithm could successfully track seven out of the eight airplanes. Based on the tracking results, movement trajectories of the airplanes and their dynamic properties were also estimated.
more »
« less
Tracking Multiple Ground Objects Using a Team of Unmanned Air Vehicles
This paper proposes a system architecture for tracking multiple ground-based objects using a team of unmanned air systems (UAS). In the architecture pipeline, video data is processed by each UAS to detect motion in the image frame. The ground-based location of the detected motion is estimated using a geolocation algorithm. The subsequent data points are then process by the recently introduced Recursive RANSAC (R-RANSASC) algorithm to produce a set of tracks. These tracks are then communicated over the network and the error in the coordinate frames between vehicles must be estimated. After the tracks have been placed in the same coordinate frame, a track-to-track association algorithm is used to determine which tracks in each camera correspond to tracks in other cameras. Associated tracks are then fused using a distributed information filter. The proposed method is demonstrated on data collected from two multi-rotors tracking a person walking on the ground.
more »
« less
- Award ID(s):
- 1650547
- NSF-PAR ID:
- 10053369
- Date Published:
- Journal Name:
- Lecture notes in control and information sciences
- Volume:
- 474
- ISSN:
- 0170-8643
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
more » « less
Abstract Magnetic anomalies over mid‐ocean ridge flanks record the history of geomagnetic field reversals, and the width of magnetized crustal blocks can be combined with absolute dates to generate a Geomagnetic Polarity Timescale (GPTS). We update here the current GPTS for the Late Cretaceous‐Eocene (chrons C33–C13, ~84–33 Ma) by extending to several spreading centers the analysis that originally assumed smoothly varying spreading rates in the South Atlantic. We assembled magnetic anomaly tracks from the southern Pacific (23 ship tracks), the northern Pacific (35), the southern Atlantic (45), and the Indian Ocean (51). Tracks were projected onto plate tectonic flow lines, and distances to magnetic polarity block boundaries were estimated by fitting measured magnetic anomalies with a Monte Carlo algorithm that iteratively changed block model distances and anomaly skewness angles. Distance data from each track were then assembled in summary sets of block model distances over 13 ridge flank regions. We obtained a final MQSD20 GPTS with another Monte Carlo algorithm that iteratively perturbed ages of polarity chron boundaries to minimize the variability of spreading rates over all ridge flanks and fit an up‐to‐date set of radioisotopic dates. The MQSD20 GPTS highlights a major plate motion change at 50–45 Ma, when spreading rates decreased in the Indian Ocean as India collided with Eurasia while spreading rates increased in the South Atlantic and Northern Pacific and the Hawaii‐Emperor seamount chain changed its orientation.
-
more » « less
ABSTRACT Nearly a hundred stellar streams have been found to date around the Milky Way and the number keeps growing at an ever faster pace. Here we present the galstreams library, a compendium of angular position, distance, proper motion, and radial velocity track data for nearly a hundred (95) Galactic stellar streams. The information published in the literature has been collated and homogenized in a consistent format and used to provide a set of features uniformly computed throughout the library: e.g. stream length, end points, mean pole, stream’s coordinate frame, polygon footprint, and pole and angular momentum tracks. We also use the information compiled to analyse the distribution of several observables across the library and to assess where the main deficiencies are found in the characterization of individual stellar streams, as a resource for future follow-up efforts. The library is intended to facilitate keeping track of new discoveries and to encourage the use of automated methods to characterize and study the ensemble of known stellar streams by serving as a starting point. The galstreams library is publicly available as a python package and served at the galstreams GitHub repository.
-
more » « less
Abstract This study presents a particle filter based framework to track cardiac surface from a time sequence of single magnetic resonance imaging (MRI) slices with the future goal of utilizing the presented framework for interventional cardiovascular magnetic resonance procedures, which rely on the accurate and online tracking of the cardiac surface from MRI data. The framework exploits a low-order parametric deformable model of the cardiac surface. A stochastic dynamic system represents the cardiac surface motion. Deformable models are employed to introduce shape prior to control the degree of the deformations. Adaptive filters are used to model complex cardiac motion in the dynamic model of the system. Particle filters are utilized to recursively estimate the current state of the system over time. The proposed method is applied to recover biventricular deformations and validated with a numerical phantom and multiple real cardiac MRI datasets. The algorithm is evaluated with multiple experiments using fixed and varying image slice planes at each time step. For the real cardiac MRI datasets, the average root-mean-square tracking errors of 2.61 mm and 3.42 mm are reported respectively for the fixed and varying image slice planes. This work serves as a proof-of-concept study for modeling and tracking the cardiac surface deformations via a low-order probabilistic model with the future goal of utilizing this method for the targeted interventional cardiac procedures under MR image guidance. For the real cardiac MRI datasets, the presented method was able to track the points-of-interests located on different sections of the cardiac surface within a precision of 3 pixels. The analyses show that the use of deformable cardiac surface tracking algorithm can pave the way for performing precise targeted intracardiac ablation procedures under MRI guidance. The main contributions of this work are twofold. First, it presents a framework for the tracking of whole cardiac surface from a time sequence of single image slices. Second, it employs adaptive filters to incorporate motion information in the tracking of nonrigid cardiac surface motion for temporal coherence.
-
more » « less
Abstract Marker‐based motion capture techniques are commonly used to measure human body kinematics. These techniques require an accurate mapping from physical marker position to model marker position. Traditional methods utilize a manual process to achieve marker positions that result in accurate tracking. In this work, we present an optimization algorithm for model marker placement to minimize marker tracking error during inverse kinematics analysis of dynamic human motion. The algorithm sequentially adjusts model marker locations in 3‐D relative to the underlying rigid segment. Inverse kinematics is performed for a dynamic motion capture trial to calculate the tracking error each time a marker position is changed. The increase or decrease of the tracking error determines the search direction and number of increments for each marker coordinate. A final marker placement for the model is reached when the total search interval size for every coordinate falls below a user‐defined threshold. Individual marker coordinates can be locked in place to prevent the algorithm from overcorrecting for data artifacts such as soft tissue artifact. This approach was used to refine model marker placements for eight able‐bodied subjects performing walking trials at three stride frequencies. Across all subjects and stride frequencies, root mean square (RMS) tracking error decreased by 38.4% and RMS tracking error variance decreased by 53.7% on average. The resulting joint kinematics were in agreement with expected values from the literature. This approach results in realistic kinematics with marker tracking errors well below accepted thresholds while removing variance in the model‐building procedure introduced by individual human tendencies.
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- The DOI is not currently available.
