Integrating Remote Sensing (Aerial Images), Geographic Information System (GIS), and Global Positioning System (GPS) Data: the Case of Mapping Weeds in a Moody County (South Dakota) Field
Ground-based weed sampling is labour intensive and expensive. However, research indicates that by understanding weed variability, management decisions and profitability can be improved. Remote sensing is a technique that provides an overview of the field. The objective of this study, conducted in Moody County, South Dakota, USA, was to develop a methodology for creating post-emergent herbicide application maps for a 65-ha field using remote sensing. Single-band level-slicing was done on a near-infrared (NIR) band collected by aerial platform on 14 June 1999 and was compared to information collected by ground-truthing over 1100 points on the same date. Areas with weeds had lower brightness values (BV) than areas without weeds. Brightness value boundaries for "treat" and "no-treat" zones were set up either by using data that contained 85% of the values from the weedy areas (80-100 BV) or using the 95% confidence interval (97-99) of the mean BV for the weedy areas. Using the BV range from 80 to 110 resulted in maps with contiguous treatment areas, and a producer accuracy of 85%. Using this criterion, about 42% of the field would not be treated and the user accuracy for "no-treat" zones was 90%. The 95% confidence interval method reduced the "treat" zone to 26% of the field, however, 84% of the areas that should have been treated were placed in the "no-treat" zone, clearly much too great an error for a producer. These data indicate that remote sensed imagery can be an important tool to locate treatment areas, but care must be taken in choosing the appropriate criteria on which to base recommendations.
Proceedings of the 5th International Conference on Precision Agriculture, Bloomington, Minnesota, USA, 16-19 July, 2000
Thanapura, P.; Clay, S. A.; Clay, D. E.; Cole, C.; Dalsted, K.; and O'Neill, M, "Integrating Remote Sensing (Aerial Images), Geographic Information System (GIS), and Global Positioning System (GPS) Data: the Case of Mapping Weeds in a Moody County (South Dakota) Field" (2000). Agronomy, Horticulture and Plant Science Faculty Publications. 193.