A method was presented to transform sampled machine positional data obtained from a global positioning system (GPS) receiver into a two-dimensional raster map of number of passes as a function of location. The effect of three sources of error in the transformation process were investigated: path sampling rate (receiver sampling fre - quency); output raster resolution; and GPS receiver errors. Total accuracy of traffic maps across a site (the summed ar - eas receiving one, two, three, etc. passes) was not greatly affected by the error sources. The estimate of number of passes at a specific point, however, was heavily dependent on the presence of errors in the input data. Adding random offsets to each GPS position, for example, resulted in less than a 35% chance that an individual pixel would be classi - fied correctly following transformation when compared with a reference raster. Although the absolute accuracy of the GPS-transformation system was not defined, it was concluded that data derived from applying it could be used to make estimates of total site disturbance and to identify regions of higher or lower disturbance but was less effective when applied in defining number of passes at a given point in a stand. Resume : Les auteurs presentent une methode de traitement de donnees GPS qui permet de convertir un echantillon de points de positionnement d'un equipement forestier, en une couverture matricielle relatant le nombre de passages selon la localisation. L'effet de trois sources d'erreurs dans le processus de transformation a ete etudie : la frequence d'echantillonnage du recepteur, la definition des pixels et l'erreur de reception du GPS. La precision de la cartographie des passages sur un site (sommation des surfaces ayant recu un, deux, trois, … passages) n'est pas significativement affectee par ces sources d'erreurs. Toutefois, l'estimation du nombre de passages en un point specifique est grandement influencee par la presence d'erreurs liees aux donnees sources. L'ajout de points aleatoires a chacune des positions GPS permet, par exemple, de reduire a moins de 35% la probabilite qu'un pixel soit classe correctement suite au traite- ment, comparativement a une reference matricielle. Bien que la precision absolue du systeme « GPS-transformation » n'ait pas ete determinee, les auteurs concluent que l'information extraite par cette methode peut etre utilisee pour esti- mer la perturbation infligee aux sites et identifier les sections de faible et de haute perturbations. La procedure s'est cependant averee moins efficace pour definir le nombre de passages en un point donne, a l'interieur d'un peuplement.
[1]
Ralph A. Palmer.
American society of agricultural engineers
,
1925,
Journal of the A.I.E.E..
[2]
E. Carter,et al.
ASSESSMENT OF SOIL STRENGTH VARIABILITY IN A HARVESTED LOBLOLLY PINE PLANTATION IN THE PIEDMONT REGION OF ALABAMA, UNITED STATES*
,
2000
.
[3]
Timothy P. McDonald,et al.
Accuracy of Tracking Forest Machines with GPS
,
2000
.
[4]
R. Sands,et al.
Compaction of forest soils. A review
,
1980
.
[5]
Pacific Northwest Cooperative Extension,et al.
Compaction of forest soils
,
1981
.
[6]
Bryce J. Stokes,et al.
Comparison of two thinning systems. Part 1. Stand and site impacts
,
1995
.
[7]
J. L. Smith,et al.
Using classification error matrices to improve the accuracy of weighted land-cover models
,
1987
.
[8]
I. Wästerlund.
Extent and causes of site damage due to forestry traffic
,
1992
.
[9]
E. Carter,et al.
Relationship Between Site Disturbance and Forest Harvesting Equipment Traffic
,
1998
.
[10]
S. McMahon,et al.
Accuracy of Two Ground Survey Methods for Assessing Site Disturbance
,
1995
.