Academic journal article URISA Journal

Comparing GPS Receivers: A Field Study

Academic journal article URISA Journal

Comparing GPS Receivers: A Field Study

Article excerpt

Introduction

The use of GPS receivers has become widespread over recent years. Many applications, from hunting to surveying, benefit greatly from these devices. The level of accuracy required from application to application varies greatly. It is important to recognize the grades of GPS receivers, namely consumer, mapping, and survey grade, and their ability to accurately map features with or without differential correction. The accuracies of these receivers range from centimeter to several meters, making it necessary to evaluate how accuracy and precision can affect individual applications.

When using a GPS receiver to collect field data, accuracy can be very important, especially when collecting data for use with high-spatial resolution imagery. Quickbird multispectral imagery, for example, achieves a resolution of 2.4 meters per pixel. To coregister corresponding ground sample locations within the correct pixel(s), an accurate GPS receiver is required. To ensure that each field observation is coregistered with the correct pixel, a GPS receiver must achieve an accuracy < 50 percent of the pixel size (e.g., +/- 1.2 m @ 95 percent CI where Quickbird imagery will be used). The increased availability of less expensive, consumer-grade GPS receivers, such as the HP/Pharos receiver used in this study, that are compatible with common GPS software, such as ESRI's ArcPad or Trimble's TerraSync, has raised concern about data quality. Many such receivers collect data that cannot be differentially corrected, increasing the margin of positional errors in the data collected. Consumer-grade receivers are also unable to control the quality of PDOP during data collection, further increasing positional error. To assess the validity of these concerns a field study was designed to calculate and compare the accuracy and precision of several GPS receivers. The goal of this study was to identify the receivers most appropriate for various research, remote sensing, and GIS applications.

Similar studies have been conducted in which GPS receiver accuracy has been investigated. Some studies compared receivers under various collection protocols. Studies conducted in Ridley State Park in Pennsylvania (McCullough 2002) and the Clackamas Test Network in Oregon (Chamberlain 2002) tested the capability of the Trimble GeoXT receiver in forested and clear areas with similar procedures and yielding comparable results in each study. Using internal and external receivers (antenna located within the receiver--internal, antenna attached externally to receiver--external), the studies experimented with WAAS and postprocess differential correction techniques, but used higher PDOP masks (e.g., PDOP mask = 7.0) than used in this study (PDOP mask = 5.0). Published studies comparing various GPS receivers are limited. One study completed in the summer of 2000 compared the accuracy of five different GPS receivers under forest canopy cover with Selective Availability (SA) off (Karsky et al. 2000). In this study, WAAS was not used because it was not yet available. Differential correction was performed on files that could be corrected and positions were taken at known points in forested areas with 1, 60, and 120 positions averaged for each point. None of the above studies mentions how often points were collected over time or how many times points were collected. Each study concluded the receiver tested was appropriate for its research purposes, whatever those may have been. Overall, previous studies have taken into account some of the aspects related to GPS receiver accuracy, but a comprehensive analysis was not completed.

A study conducted in McDonald Forest, located in western Oregon, investigated the accuracy and reliability of consumer-grade GPS receivers under differing canopy conditions. Six different GPS receivers were evaluated for accuracy under three different canopies: open sky, young forest, and closed canopy. Although the collected data was unable to be differentially corrected, points were averaged and compared relative to the known location, allowing for the receivers' accuracies to be compared to one another (Wing et al. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.