|
|
|
Statistical Object Identification, Tracking and Analysis
|
|
|
| Scientists are overwhelmed by the vast amounts of data available from instruments, experiments, and simulations; all information technology trends point to a continuing increase in the size of these
datasets. Unfortunately, the ability to collect, store, and manipulate ever-greater volumes of data will not translate into greater scientific understanding without development of suitable software. One way to conceptualize the process of discovering patterns in these large spatiotemporal datasets is to
identify objects, which we take to be spatially coherent regions of interest that evolve through time. This provides one way for scientists to rise above a pixel-level understanding of their data to a more natural and physically relevant description in terms of salient features.
To analyze objects, they must be identified in individual images, tracked through image sequences, and their temporal behaviors modeled quantitatively. We may solve a large class of object-identification problems with modern, general-purpose spatial statistics methods which employ unsupervised learning or scientist-provided labelings to define local assessments of activity. These pixel-level cues are progressively linked into a coherent, object-level scene description. Similarly general statistical time series models are used for object tracking and analysis. Expressing the models in a neutral, mathematical/statistical language allows repeatability, model refinement over time, and model exchange among investigators. Our use of open standards for structured data (e.g., object models and datasets in XML with suitable schemata) ensures portability across applications and future extensibility. The system is field-tested on a multiwavelength series of solar images in conjunction with solar physicists. Strong collaborations with domain scientists ensure successful deployment into other Sun-Earth connection missions, and OSS application areas like planetary science and geophysics. |
Bibliography
| | Study of differences between sunspot area data determined from ground-based and space-borne observations
Turmon, M.; Gyori, L.; Baranyi, T.; Pap, J. M., Adv. Space Research, (2004) 34 pp. 269-273
| |
| Symmetric Normal Mixtures
Turmon, M.; Proc. Compstat-2004, (2004)
| |
| Comparison of image-processing methods to extract sunspots
Turmon, M.; Gyori, L.; Baranyi, T.; Pap, J. M., Proc. SoHO-11 Workshop, (2002) ESA SP-508 pp. 203-208
| |
| Study of the SOHO/VIRGO irradiance variations using MDI and Kitt Peak images
Turmon, M; J. M. Pap; H. Jones; L. Floyd, Proc. SoHO-11 Workshop, (2002) ESA SP-508
| |
| Statistical Pattern Recognition for Labeling Solar Active Regions: Application to SOHO/MDI Imagery
Turmon; Pap; Mukhtar; Astrophysical Journal, (2002) 568 pp. 396-407
| |
|
|
