A Method For Relating GPS Performance To Forest Canopy
Authors
N. M. Holden
University College Dublin, Ireland
A. A. Martin
University College Dublin, Ireland; Dublin Institute of Technology, Ireland
P. M.O. Owende
University College Dublin, Ireland
S. M. Ward
University College Dublin, Ireland
Abstract
This paper describes the development of a quantitative method of classifying forest canopy that can be related to degradation in Differential GPS (DGPS) performance. Using digital images taken vertically skywards at 20 sites within a forest on the east coast of Ireland, canopy cover was described using percentage sky obstruction (Op), largest hole (DTmax), and the fragmentation of sky view (DTp) using pixel count and distance transform data. Statistical methods were used to produce three clusters which related to canopy cover in stands of Picea sitkensis and Picea contorta. DGPS data were then collected on 10 separate dates at the 20 sites, every two seconds for five minutes, and the two-dimensional standard deviation of the position fix was used to represent DGPS precision. The standard deviations ranged from 0.5 m to 9.7 m (compared to the manufacturer's specification of approximately 1 m in the horizontal). Precision was found to be related to the total obstruction, the size of the largest hole in the canopy, and the fragmentation of the sky view. Where there was little obstruction (less than 20%) or fragmentation, the DGPS performance was effectively the same as outside the forest, open canopy caused a 2-3 fold degradation in precision, and closed canopy a 5-7 fold degradation. It is suggested that the methods presented could be used to rapidly relate DGPS performance to forest canopy, and therefore may be useful in pre-planning inventory mapping and future machine operations using guidance systems.
