Results 11 -
19 of
19
Loop Closure Transformation Estimation and Verification Using 2D LiDAR Scanners
"... Abstract — In many simultaneous localization and mapping (SLAM) systems, it is desirable to exploit the fact that the system is traversing though a previously visited environment. Once these locations, commonly known as loop closures, have been detected the system must be able to both compute and ve ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract — In many simultaneous localization and mapping (SLAM) systems, it is desirable to exploit the fact that the system is traversing though a previously visited environment. Once these locations, commonly known as loop closures, have been detected the system must be able to both compute and verify the relative transformation between proposed locations. In this paper we present two independent algorithms, using 2D LiDAR scanners, for robustly computing the transformation between arbitrary locations with overlapping geometry and validating the resulting transforms. First, a scan matching algorithm based on a genetic search and a fractional distance metric is presented. Secondly, two metrics are proposed to verify the recovered transforms. Through experimental results the proposed algorithms are shown to robustly estimate and validate loop closure transformations for both manually and automatically defined candidates. I.
Watertight Planar Surface Reconstruction of Voxel Data
, 2012
"... There are many scenarios where a 3D shape is represented by a voxel occupancy grid. Oftentimes it is desirable to convert data from this format to a triangulated mesh that represents the surface of the volume described by the occupied ..."
Abstract
- Add to MetaCart
(Show Context)
There are many scenarios where a 3D shape is represented by a voxel occupancy grid. Oftentimes it is desirable to convert data from this format to a triangulated mesh that represents the surface of the volume described by the occupied
100
"... In this paper, we propose an algorithm to automatically identify window regions on exterior facing facades of buildings using interior 3D point cloud resulting from an ambulatory backpack sensor system, outfitted with multiple LiDAR sensors and cameras. We develop a set of discriminative features fo ..."
Abstract
- Add to MetaCart
(Show Context)
In this paper, we propose an algorithm to automatically identify window regions on exterior facing facades of buildings using interior 3D point cloud resulting from an ambulatory backpack sensor system, outfitted with multiple LiDAR sensors and cameras. We develop a set of discriminative features for the task, namely visual brightness, infrared opaqueness, and an occlusion indicator, within a Markov Random Field (MRF) framework to provide structured prediction for window or glass regions. A preprocessing classifier is trained on the features to produce node potentials, and large margin parameter training is used to boost performance. Our algorithm has been trained on data taken at a university building with a total façade area of 637 m 2, and has been tested on walls in a Walgreens store, an office building in San Francisco, and a hotel in Houston, with a total exterior façade area of 405 m 2. Window regions are successfully identified with 85.2 % F1-score and 82.5% accuracy. 1.
Energy Minimization
"... In this paper, we propose a novel surface completion method to generate plausible shapes and textures for missing regions of 3D models. The missing regions are filled in by minimizing two energy functions for shape and texture, which are both based on similarities between the missing region and the ..."
Abstract
- Add to MetaCart
(Show Context)
In this paper, we propose a novel surface completion method to generate plausible shapes and textures for missing regions of 3D models. The missing regions are filled in by minimizing two energy functions for shape and texture, which are both based on similarities between the missing region and the rest of the object; in doing so, we take into account the positive correlation between shape and texture. We demonstrate the effectiveness of the proposed method experimentally by applying it to two models.
TEXTURING LONG PLANAR SURFACES WITH IMPRECISE CAMERA POSES FOR INDOOR 3D MODELING
"... Automated 3D modeling of building interiors is useful in applications such as virtual reality and environment mapping. Texture mapping walls is an important step in visualizing the results of an indoor 3D modeling system. Methods to localize the camera in the 3D scene often are not pixel accurate, m ..."
Abstract
- Add to MetaCart
(Show Context)
Automated 3D modeling of building interiors is useful in applications such as virtual reality and environment mapping. Texture mapping walls is an important step in visualizing the results of an indoor 3D modeling system. Methods to localize the camera in the 3D scene often are not pixel accurate, meaning that when multiple images are used for texture mapping there are seams and discontinuities between these images. Several approaches to this problem have been proposed but each suffer from a distinct problem of error accumulation for long chains of images, such as those from a long corridor. We propose a new approach to texture mapping planar surfaces that eliminates discontinuities between images but does not suffer from error accumulation for long chains. We validate this approach using images from several long hallways with data generated by a human operated backpack 3D indoor modeling system. Index Terms — 3D modeling, mosaicing, texture mapping 1.
unknown title
"... In this paper, we propose an algorithm to automatically identify window regions on exterior facing facades of buildings using interior 3D point cloud resulting from an ambulatory backpack sensor system, outfitted with multiple LiDAR sensors and cameras. We develop a set of discriminative features fo ..."
Abstract
- Add to MetaCart
(Show Context)
In this paper, we propose an algorithm to automatically identify window regions on exterior facing facades of buildings using interior 3D point cloud resulting from an ambulatory backpack sensor system, outfitted with multiple LiDAR sensors and cameras. We develop a set of discriminative features for the task, namely visual brightness, infrared opaqueness, and an occlusion indicator, within a Markov Random Field (MRF) framework to provide structured prediction for window or glass regions. A preprocessing classifier is trained on the features to produce node potentials, and large margin parameter training is used to boost performance. Our algorithm has been trained on data taken at the 3rd floor of Cory Hall at UC Berkeley, with a total façade area of 269.1 m2, and has been tested on walls taken on the 2nd floor of Cory Hall, a Walgreens, and an office building in San Francisco, with a total exterior façade area of 454.6 m2. Window regions are successfully identified with 85.5% F1-score and 94.2 % accuracy. 1.
Indoor WiFi Localization with a Dense Fingerprint Model
"... Abstract—WiFi-based localization is a popular approach for positioning a WiFi-enabled device in an indoor environment. Most implementations rely on querying fingerprint databases, created by stop and go sampling of WiFi signals at discrete locations used as reference points. In this paper, we propo ..."
Abstract
- Add to MetaCart
Abstract—WiFi-based localization is a popular approach for positioning a WiFi-enabled device in an indoor environment. Most implementations rely on querying fingerprint databases, created by stop and go sampling of WiFi signals at discrete locations used as reference points. In this paper, we propose an approach for rapid creation of a dense WiFi fingerprint database using a human operated ambulatory backpack and a single walkthrough and data collect in an indoor environment. In addition, we present and compare the performance of 4 algorithms for localizing mobile devices based on the collected fingerprints that take advantage of the dense database. We show that it is possible to achieve mean error of 2.8 meters with 90th percentile of 5.0 meters using one of our algorithms. Keywords—indoor localization; WiFi localization; WiFi fingerprints;
a postdoctoral fellow, Jason Cramer is an undergraduate student, and Avideh Zakhor is a professor. Automatic Generation of 3D Thermal Maps of Building Interiors
"... Most existing approaches to characterizing thermal properties of buildings and heat emissions from their elements rely on manual inspection and as such are slow, and labor intensive. This is often a daunting task, which requires several days of on-site inspection. In this paper, we propose a fully a ..."
Abstract
- Add to MetaCart
(Show Context)
Most existing approaches to characterizing thermal properties of buildings and heat emissions from their elements rely on manual inspection and as such are slow, and labor intensive. This is often a daunting task, which requires several days of on-site inspection. In this paper, we propose a fully automatic approach to construct a 3D thermal point cloud of the building interior reflecting the geometry including walls, floors, and ceilings, as well as structures such as furniture, lights, windows, and plug loads. Our approach is based on a wearable ambulatory backpack comprising multiple sensors such as Light Detection And Ranging (LiDAR) scanners, and Infrared and optical cameras. As the operator wearing the backpack walks through the building, the LiDAR scans are collected and processed in order to compute the 3D geometry of the building. Furthermore, the Infrared cameras are calibrated intrinsically and extrinsically such that the captured images are registered to the captured geometry. Thus, the temperature data in the Infrared images is associated with the geometry resulting in a “thermal 3D point cloud”. The same process can be repeated using optical imagery resulting in a “visible 3D point cloud”. By visualizing the two point clouds simultaneously in interactive rendering tools, we can virtually walk through the thermal and optical 3D point clouds, toggle between them, identify and annotate, “hot ” regions, objects, plug loads, thermal and moisture leaks, and document their location with fine spatial granularity in the 3D point clouds.
Texture Mapping 3D Models of Indoor Environments with Noisy Camera Poses
"... Automated 3D modeling of building interiors is used in applications such as virtual reality and environment mapping. Texturing these models allows for photo-realistic visualizations of the data collected by such modeling systems. While data acquisition times for mobile mapping systems are considerab ..."
Abstract
- Add to MetaCart
Automated 3D modeling of building interiors is used in applications such as virtual reality and environment mapping. Texturing these models allows for photo-realistic visualizations of the data collected by such modeling systems. While data acquisition times for mobile mapping systems are considerably shorter than for static ones, their recovered camera poses often suffer from inaccuracies, resulting in visible discontinuities when successive images are projected onto a surface for texturing. We present a method for texture mapping models of indoor environments that starts by selecting images whose camera poses are well-aligned in two dimensions. We then align images to geometry as well as to each other, producing visually consistent textures even in the presence of inaccurate surface geometry and noisy camera poses. Images are then composited into a final texture mosaic and projected onto surface geometry for visualization. The effectiveness of the proposed method is demonstrated on a number of different indoor environments.
