Anomaly Density Estimation from Strip Transect Data: Pueblo of Isleta Example

Abstract

Extensive geophysical surveying done at a wide variety of sites across the country has shown that on any one site, the density, or intensity, of geophysical anomalies varies from one location to another. This spatial variation makes it difficult to determine the total number of anomalies at any one location given that it may not be possible to geophysically survey the entire site. However, it is desirable to accurately estimate the number of anomalies at specific locations in order to determine boundaries of target locations or other regions where the anomaly density is above the background density. Also, an accurate picture of the anomaly distribution is required for pre excavation cost estimation and planning worker safety measures. This report documents an approach to estimating spatially variable anomaly density from limited transect data. This approach is then tested using a data set collected at the Pueblo of Isleta S1 site in central New Mexico. The data set was collected by the Naval Research Laboratory using an airborne magnetometer array (NRL, 2005). The data set provides an exhaustive picture of the anomaly locations over an area of nearly 1500 acres at the Isleta site. Excavation at this site has identified UXO in several locations and the entire 1500 acres may be considered as a target area. This study on anomaly density estimation does not attempt to classify the site into different types of areas (e.g., background vs. target area) nor establish boundaries for potential target areas. The work here is solely focused on estimating the number of anomalies at all locations. Three different transect sampling designs covering 2.85, 0.90 and 0.45 percent of the total site are used to provide the data for the estimation. The results of the density estimation are evaluated by comparing the estimates to the true surveyed anomalies at the unsampled locations.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 01, 2005
Accession Number
ADA606782

Entities

People

  • Brent Pulsipher
  • Sean A. Mckenna

Organizations

  • Pacific Northwest National Laboratory

Tags

DTIC Thesaurus Topics

  • Airborne
  • Algorithms
  • Boundaries
  • Cell Size
  • Coordinate Systems
  • Cost Estimates
  • Data Sets
  • Detection
  • Errors
  • Excavation
  • Grids
  • Military Research
  • New Mexico
  • Spatial Distribution
  • Statistics
  • Surveys
  • Visual Inspection

Readers

  • Computer Vision.
  • Military/Explosive Ordnance Disposal (EOD) Technology
  • Statistical inference.