ARTISAN: An Integrated Scene Mapping and Object Recognition System
نویسندگان
چکیده
Integration of three-dimensional textured scene mapping and object recognition presents many opportunities for assisted automation. We present Artisan, a software package that synthesizes these elements to form a user-friendly whole. Artisan uses a variety of 3D sensors, including laser range scanners and stereo systems, to acquire both image and range data. Artisan automatically finds the transformations between data taken at multiple sensor viewpoints using matching algorithms. The data from these viewpoints are then merged together to form an integrated textured map of the entire scene. Other user or sensor input can also inserted into the scene. Using object recognition with an expandable library of objects, Artisan can identify and locate simple and complex scene features. With this identity and transformation information, it is able to support many operations, including semi-automatic robotic teleoperation and navigation. After mapping and recognition, the identity, position, and orientation of the objects in the scene can be automatically transferred from the Artisan system into other software, including robotic teleoperation packages. Numerous opportunities for automation exist during the operations stage as a result of this increased world knowledge. This work was performed under contracts DE-AR26-97FT34314 and DE-AC21-92MC29104 for the Department of Energy, Federal Energy Technology Center, Morgantown, West Virginia.
منابع مشابه
Scene and Object Recognition with Supervised Nonlinear Neighborhood Embedding
Image category recognition is important to access visual information on the level of objects and scene types. In this paper, we develop a Supervised Nonlinear Neighborhood Embedding (SNNE) subspace algorithm of different visual features for object and scene recognition, which learns an adaptive nonlinear subspace by preserving the neighborhood structure of the visual feature space. In the propo...
متن کامل2D/3D Object Categorization for Task Based Grasping
Fig. 1. System outline. First row: Data acquisition using ARMAR III robot head and view of a typical experimental scene. Second row: Segmented objects in the same scene. Third row: Integrated 2D and 3D Object Categorization Systems (OCSs). Fourth row: Generation of grasping points by Bayesian network. Fifth row: Experimental scene with grasping points for categorized objects and desired task (t...
متن کاملBasic level scene understanding: from labels to structure and beyond Citation
An early goal of computer vision was to build a system that could automatically understand a 3D scene just by looking. This requires not only the ability to extract 3D information from image information alone, but also to handle the large variety of different environments that comprise our visual world. This paper summarizes our recent efforts toward these goals. First, we describe the SUN data...
متن کاملImprovement of Navigation Accuracy using Tightly Coupled Kalman Filter
In this paper, a mechanism is designed for integration of inertial navigation system information (INS) and global positioning system information (GPS). In this type of system a series of mathematical and filtering algorithms with Tightly Coupled techniques with several objectives such as application of integrated navigation algorithms, precise calculation of flying object position, speed and at...
متن کاملMulti-Modal Scene Interpretation
The visionary goal of developing an easy to use service robot implies several key tasks such as speech understanding, object recognition and scene understanding. Besides the more sensor-oriented capabilities such systems need extensive meta knowledge, e.g., about mental representations of spatial relations to match the view between man and machine. Only if all parts fit together an unrestricted...
متن کامل