|United States Department of Agriculture
Tests to evaluate the "walk method" of traversing, used to determine the area of a polygon, were conducted at the GPS Hardwood Test Site near Bedford, IN, with the Trimble ProXL GPS receiver. The results compare the accuracy of the area obtained using the walk method of traversing with the known area of the polygon. The time required to traverse the polygon was also obtained.
The GPS Hardwood Test Site is on the Hoosier National Forest near Bedford, IN. It is on gentle terrain under a dense canopy of uneven-aged oak, beech, and hickory trees typical of eastern hardwood forests, with canopy tops 31 to 37 meters (100 to 120 feet) above the ground. The test course has seven turning points located on two finger ridges with one point in the bottom of the draw between the ridges. The polygon defined by the seven turning points contains 3.32 acres. All turning points have been located to second-order class 1 accuracy or better.
A Trimble ProXL eight-channel GPS receiver (Figure 1) with an MC-V data logger was used to collect the data. It was set up for the manual 3-D mode of operation, with a 5-second data logging interval, an SNR and PDOP mask of 6, and an elevation mask of 15 degrees. Differential data was provided from the Bedford Community Base Station (CBS), which is about 17 km (11 miles) from the test site. It was set up to collect synchronous data at the 5-second logging interval, with an SNR and PDOP mask set at 6 and an elevation mask of 10 degrees.
|Figure 1--The Trimble ProXL eight-channel GPS receiver.|
Data for the walk method were collected by starting at one turning point and walking in a straight line to the next turning point, then to the next one until the perimeter of the polygon had been traversed (closing the polygon at the starting point).
The data logging interval was set at 5 seconds. When the signal was lost while traversing, the pace was slowed until the signal was reacquired and data were again being collected. Care was taken at each turning point to make sure data were being collected. In some cases, this required slowing down or even stopping for a few seconds. The time required to traverse the course using this method was obtained from the data file for each run and is shown in Table 1. It appeared to require less time than the "point method," which requires 3 minutes of data collection (under ideal conditions) at each turning point. All GPS data were differentially corrected using Trimble PFINDER MCORR400 software with data from the Bedford CBS. The results are shown in Table 1.
Table 1--Results of GPS area measurements using the "walk method" of traversing and differentially corrected data from tests conducted under the canopy at the Bedford, IN, GPS hardwood test site. The Trimble ProXL GPS receiver was set for manual 3-D operation, a 5-second data logging interval, SNR and PDOP Mask of 6, and an elevation mask of 15 degrees.
This data indicates an average error of only 1.7% when using the walk method of traversing to determine the area of the 3.32-acre polygon under a dense hardwood canopy. All area measurements were within 3.3% (0.11 acres) of the actual area. The reason for the 3.3% error in area on the first and last runs of the test is unknown. The data were collected on different days and at different times of the day. Accuracy expressed as a percent of the area would be expected to increase with larger areas. The average time to collect the data was 17.4 minutes with a maximum of 31.6 minutes. The time required to collect the data varies greatly because the dense canopy weakens the satellite signal. The walk method is more efficient than the "point method," which under ideal conditions would require 3 minutes to collect 180 position records at each of the seven turning points, or 21 minutes. However, under a dense canopy, collecting the 180 position records takes a lot longer than 3 minutes at each station.
Tony Jasumback is a mechanical engineer and Project Leader at MTDC.
Doug Luepke is a photogrammetrist for the Southern Region in Atlanta, GA.
Rich McCullough is a forester for the Northeastern Forest Experiment Station in Radnor, PA.
Dale Weigel is a forester for the North Central Forest Experiment Station in Bedford, IN.
For further technical information, contact Tony Jasumback at MTDC.Phone: (406) 329-3922 Fax: (406) 329-3719 Email Document Request