Learn More
This paper presents a new state-of-the-art for document image classification and retrieval, using features learned by deep convolutional neural networks (CNNs). In object and scene analysis, deep neural nets are capable of learning a hierarchical chain of abstraction from pixel inputs to concise and descriptive representations. The current work explores(More)
Augmented Reality (AR) combines a live view of a real world environment with computer-generated virtual content. An AR system poses unique challenges including requiring a high quality camera pose estimate and operating on resource-limited platforms. We present a full system based on a hybrid approach using ORB binary features and optic flow that is able to(More)
The use of unmanned aerial vehicles (UAVs) has the potential to significantly improve the situation awareness of emergency first responders working at urban disaster sites. Having the characteristics of being small, light-weight and quickly deployable, UAVs offer the ability to fly over an urban disaster and provide intelligence to Urban Search and Rescue(More)
In this paper we discuss several methods for the creation of 3D models that can provide additional information to robot operators in order to improve their situation awareness of the robot being teleoperated. We derive the 3D models from spatial data gathered from an inexpensive, readily available, video game sensor. In addition, the paper introduces a new(More)
Using Visual Odometry a robot can track its trajectory using video input. This allows more accurate ego-motion estimation when compared to classical odometry which relies on measurement of wheel motion. The Microsoft Kinect sensor provides 3D imagery, similar to a LASER or LIDAR scanner, which can be used for visual odometry with a single sensor. This diers(More)
Various natural and human-made events can occur in urban settings resulting in buildings collapsing and trapping victims. The task of a structural engineer is to survey the resulting rubble to assess its safety and arrange for structural stabilization, where necessary. Urban Search and Rescue (USAR) operations can then begin to locate and rescue people. Our(More)
A 3D map of the interior of a disaster site that pinpoints the location of trapped victims would greatly aid search and rescue efforts. We propose using a canine-mounted RGB-D sensor; a trained rescue dog can carry an image sensor through the site to build a 3D model useful for rescuers. However, the registration of the data provides challenges beyond those(More)
A Destroyed Environment (DE) is created by disastrous events in the built environment. DEs typically consist of the structures created from the rubble of collapsed buildings-forming a chaotic, unplanned and unmapped environment in which emergency first responders must find the surviving occupants who may now be trapped and hidden within. The more knowledge(More)
Under rubble communication is a well-known difficult problem. This is due to a number of inherent difficulties including plentiful and diverse physical obstacles that challenge both wired and wireless communication. Our previous work has shown that it is possible to use Urban Search and Rescue (USAR) dogs to deliver emergency supplies, robots and(More)
  • 1