This paper presents the innovative SatTel framework, designed to automatically access, collate, process, and exploit commercial satellite imagery from a wide variety of vendors. Established vendors such as DigitalGlobe provide high resolution imagery with limited coverage, while disruptive vendors such as Planet and BlackSky provide low resolution imagery with near global coverage. SatTel provides a single point of entry for exploitation of these contrasting and complementary vendor capabilities. The authors illustrate the value of the SatTel framework via demonstrative change detection capabilities. SatTel change detection from small satellite imagery based on comparison of image to image appearance achieves mean average precision (MAP) above 0.75 for many sites compared to ground truth analyst annotation. SatTel change detection from high resolution satellite imagery based on multidimensional geometric structures achieves an average precision of 0.84 for elevation changes above 3.0 meters compared to ground truth analyst annotation.

Human pose estimation - the process of recognizing a human's limb positions and orientations in a video - has many important applications including surveillance, diagnosis of movement disorders, and computer animation. While deep learning has lead to great advances in 2D and 3D pose estimation from single video sources, the problem of estimating 3D human pose from multiple video sensors with overlapping fields of view has received less attention. When the application allows use of multiple cameras, 3D human pose estimates may be greatly improved through fusion of multi-view pose estimates and observation of limbs that are fully or partially occluded in some views. Past approaches to multi-view 3D pose estimation have used probabilistic graphical models to reason over constraints, including per-image pose estimates, temporal smoothness, and limb length. In this paper, we present a pipeline for multi-view 3D pose estimation of multiple individuals which combines a state-of-art 2D pose detector with a factor graph of 3D limb constraints optimized with belief propagation. We evaluate our results on the TUM-Campus and Shelf datasets for multi-person 3D pose estimation and show that our system significantly out-performs the previous state-of-the-art with a simpler model of limb dependency.

Multi-megapixel cameras are transforming airborne video surveillance by enabling persistent imaging of extremely large areas while providing sufficient pixel density to resolve both vehicles and pedestrians. The sheer spatial and temporal volume of data has rendered human scanning of expansive images for miniscule moving objects intractable, underscoring the importance of automated detection and tracking systems. Existing algorithms, however, are generally designed for stationary cameras and moderately-sized objects. This paper presents one of the first systems for reliably detecting and tracking low-resolution objects of varying size and shape in challenging wide-area aerial video. Significant contributions include a simple, fast approach for robust motion detection with parallax handling; spatial-temporal filtering for quickly discarding spurious detections; adaptive shape learning for unusually-shaped objects; and multi-cue fusion for state evolution that enables tracking through confusion, occlusions, and stops. Our system is highly efficient and parallelizable, processing 1-megapixel image streams in real time on a single CPU core. Experiments on a variety of data sets demonstrate that the system outperforms more traditional detection and tracking approaches, and is able to find pedestrians missed by human ground truthers despite tiny size, poor contrast, and surrounding clutter.1

The increasing availability of very high resolution satellite imagery has spurred interest in automatically detecting very fine detailed changes in an area over time, a particularly useful tool for analyzing activity in dense urban areas. However, attempting automated change detection at this resolution is difficult due to the motion parallax of elevated structures. This paper presents a comprehensive solution to change detection in areas of significant 3D relief using a new framework called volumetric appearance modeling (VAM). This approach can manage the complications of unknown and changing world surfaces by maintaining a 3D voxel-based model, where probability distributions for surface occupancy and image appearance are stored in each voxel. These distributions are continuously updated as new images are received using an adaptive learning procedure. This representation is demonstrated to produce accurate change detection results under conditions of variable illumination and viewpoint as well as haze conditions present in satellite imagery. The volumetric representation also supports automatic sensor model correction to align incoming imagery to a common geographic reference. This registration approach is demonstrated to achieve geo-positioning accuracy on the order of the ground sampling distance (GSD) or better.

The problem of detecting changes in images taken by a stationary camera has been well studied and many algorithms now exist which perform robustly in real-world applications. However no comprehensive solution exists for the analogous “3-d” change detection problem, where the camera taking the images is allowed to move and the scene is not planar. This situation is common in a growing number of applications, especially in aerial surveillance where the recent explosion of available aerial imagery has made manual examination by analysts infeasible. This thesis presents the first comprehensive solution to the 3-d change detection problem, using a new framework called volumetric appearance modeling (VAM). This approach can manage the complications of unknown and sometimes changing world surfaces by maintaining a 3-d voxel-based model, where probability distributions for surface occupancy and image appearance are stored in each voxel. These distributions are continuously updated as new images are received using an adaptive learning procedure. VAM is demonstrated to perform well on several aerial image sequences taken from helicopter from a wide range of viewpoints. The core VAM framework is extended to additionally deal with the variable lighting and haze conditions present in satellite imagery. The success of this complete framework is demonstrated on selected scenes of Baghdad, Iraq exhibiting a challenging range of viewpoints, lighting directions, and haze.

This paper examines the problem of detecting changes in a 3-d scene from a sequence of images, taken by cameras with arbitrary but known pose. No prior knowledge of the state of normal appearance and geometry of object surfaces is assumed, and abnormal changes can occur in any image of the sequence. To the authors' knowledge, this paper is the first to address the change detection problem in such a general framework. Existing change detection algorithms that exploit multiple image viewpoints typically can detect only motion changes or assume a planar world geometry which cannot cope effectively with appearance changes due to occlusion and un-modeled 3-d scene geometry (ego-motion parallax). The approach presented here can manage the complications of unknown and sometimes changing world surfaces by maintaining a 3-d voxel-based model, where probability distributions for surface occupancy and image appearance are stored in each voxel. The probability distributions at each voxel are continuously updated as new images are received. The key question of convergence of this joint estimation problem is answered by a formal proof based on realistic assumptions about the nature of real world scenes. A series of experiments are presented that evaluate change detection accuracy under laboratory-controlled conditions as well as aerial reconnaissance scenarios.-