Abstract
The effectiveness of mask setting for point positioning in forests was studied. Elevation mask did not improve the positioning accuracy in forests, because the recommended elevation mask in a treeless area was not appropriate for positioning in forests with a large stand basal area. While usual SNR mask did not improve the positioning accuracy in forests, there was a significant relation between mean SNR and positioning accuracy. Although PDOP mask reduced the positioning success rate in forests, it had the greatest effect on reducing error in the successful cases. Multiple regression analysis was applied to find factors affecting the positioning accuracy. Elevation angles of four satellites significantly affected the positioning accuracy in a treeless area, and there was significance in PDOP and mean SNR inside forests. The maximum multiple correlation coefficient was 0.45. As a result of mask setting using the chosen variables, mean positioning error was reduced by 40% and 35% for a treeless area and a forest, respectively. In conclusion, the use of PDOP and mean SNR can enable us to control the positioning accuracy in forests.
