Applications of a geographic information system and remote sensing to the identification of agricultural fields.

摘要

This thesis focuses on the use of remote sensing and a geographic information system to identify large agricultural fields and to verify Census data for western Ontario, Canada. A new method for locating at least three large agricultural fields (area larger than 14.5 ha in size) around main road intersections was developed, and the results showed that this method was accurate and efficient. The results are to be used for ground truthing and calibration of satellite imagery for monitoring agroecosystems. The areas of agriculture from image analysis were compared with Census data for the study area on a county level, and the reasons for the differences between these two databases were discussed.; Different resolutions of images were also used to evaluate the effects on identifying large agricultural fields for a 5 x 10 km test area. The results for the test area from two different resolutions of images showed that 25 m spatial resolution of Landsat TM met the requirements of our study, and high resolution of images (1 m and 4 m spatial resolutions) produced a very satisfactory result, but the costs and time consumption were higher than that of low resolution images.